diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,14677 @@ +{ + "best_metric": 0.37647188, + "best_model_checkpoint": "/mnt/petrelfs/caoyuhang/InternLM-XComposer/finetune_audio/output/sft-continue_base_silence/qwen2-audio-7b-instruct/v0-20241120-155458/checkpoint-20000", + "epoch": 3.0, + "eval_steps": 2000, + "global_step": 32220, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 9.310986964618249e-05, + "grad_norm": 1.40625, + "learning_rate": 1.2414649286157668e-07, + "loss": 0.18691748, + "memory(GiB)": 26.35, + "step": 1, + "train_speed(iter/s)": 0.046198 + }, + { + "epoch": 0.00186219739292365, + "grad_norm": 1.59375, + "learning_rate": 2.4829298572315337e-06, + "loss": 0.2265442, + "memory(GiB)": 59.78, + "step": 20, + "train_speed(iter/s)": 0.434773 + }, + { + "epoch": 0.0037243947858473, + "grad_norm": 1.4140625, + "learning_rate": 4.965859714463067e-06, + "loss": 0.21457853, + "memory(GiB)": 59.78, + "step": 40, + "train_speed(iter/s)": 0.569546 + }, + { + "epoch": 0.00558659217877095, + "grad_norm": 1.5, + "learning_rate": 7.4487895716946e-06, + "loss": 0.2096719, + "memory(GiB)": 75.39, + "step": 60, + "train_speed(iter/s)": 0.627998 + }, + { + "epoch": 0.0074487895716946, + "grad_norm": 1.3046875, + "learning_rate": 9.931719428926135e-06, + "loss": 0.21570706, + "memory(GiB)": 75.39, + "step": 80, + "train_speed(iter/s)": 0.655146 + }, + { + "epoch": 0.00931098696461825, + "grad_norm": 1.9921875, + "learning_rate": 1.2414649286157666e-05, + "loss": 0.22957432, + "memory(GiB)": 75.39, + "step": 100, + "train_speed(iter/s)": 0.680076 + }, + { + "epoch": 0.0111731843575419, + "grad_norm": 1.1796875, + "learning_rate": 1.48975791433892e-05, + "loss": 0.22518969, + "memory(GiB)": 75.39, + "step": 120, + "train_speed(iter/s)": 0.695236 + }, + { + "epoch": 0.01303538175046555, + "grad_norm": 1.2265625, + "learning_rate": 1.7380509000620735e-05, + "loss": 0.20143158, + "memory(GiB)": 75.39, + "step": 140, + "train_speed(iter/s)": 0.709963 + }, + { + "epoch": 0.0148975791433892, + "grad_norm": 1.875, + "learning_rate": 1.986343885785227e-05, + "loss": 0.19077778, + "memory(GiB)": 75.39, + "step": 160, + "train_speed(iter/s)": 0.719808 + }, + { + "epoch": 0.01675977653631285, + "grad_norm": 1.390625, + "learning_rate": 2.2346368715083797e-05, + "loss": 0.2053205, + "memory(GiB)": 75.39, + "step": 180, + "train_speed(iter/s)": 0.729386 + }, + { + "epoch": 0.0186219739292365, + "grad_norm": 1.6015625, + "learning_rate": 2.482929857231533e-05, + "loss": 0.21171951, + "memory(GiB)": 75.39, + "step": 200, + "train_speed(iter/s)": 0.735649 + }, + { + "epoch": 0.020484171322160148, + "grad_norm": 1.5390625, + "learning_rate": 2.7312228429546866e-05, + "loss": 0.19139764, + "memory(GiB)": 75.39, + "step": 220, + "train_speed(iter/s)": 0.74262 + }, + { + "epoch": 0.0223463687150838, + "grad_norm": 1.421875, + "learning_rate": 2.97951582867784e-05, + "loss": 0.19102612, + "memory(GiB)": 75.39, + "step": 240, + "train_speed(iter/s)": 0.748473 + }, + { + "epoch": 0.024208566108007448, + "grad_norm": 1.6796875, + "learning_rate": 3.227808814400993e-05, + "loss": 0.22476618, + "memory(GiB)": 75.39, + "step": 260, + "train_speed(iter/s)": 0.749743 + }, + { + "epoch": 0.0260707635009311, + "grad_norm": 1.390625, + "learning_rate": 3.476101800124147e-05, + "loss": 0.20150518, + "memory(GiB)": 75.39, + "step": 280, + "train_speed(iter/s)": 0.751916 + }, + { + "epoch": 0.027932960893854747, + "grad_norm": 1.7734375, + "learning_rate": 3.7243947858473e-05, + "loss": 0.20568705, + "memory(GiB)": 75.39, + "step": 300, + "train_speed(iter/s)": 0.755403 + }, + { + "epoch": 0.0297951582867784, + "grad_norm": 1.328125, + "learning_rate": 3.972687771570454e-05, + "loss": 0.21228166, + "memory(GiB)": 75.39, + "step": 320, + "train_speed(iter/s)": 0.758702 + }, + { + "epoch": 0.03165735567970205, + "grad_norm": 1.5546875, + "learning_rate": 4.220980757293606e-05, + "loss": 0.23770952, + "memory(GiB)": 75.39, + "step": 340, + "train_speed(iter/s)": 0.760792 + }, + { + "epoch": 0.0335195530726257, + "grad_norm": 1.7265625, + "learning_rate": 4.4692737430167594e-05, + "loss": 0.1984645, + "memory(GiB)": 75.39, + "step": 360, + "train_speed(iter/s)": 0.761537 + }, + { + "epoch": 0.035381750465549346, + "grad_norm": 1.34375, + "learning_rate": 4.717566728739913e-05, + "loss": 0.20803604, + "memory(GiB)": 45.23, + "step": 380, + "train_speed(iter/s)": 0.762314 + }, + { + "epoch": 0.037243947858473, + "grad_norm": 1.1328125, + "learning_rate": 4.965859714463066e-05, + "loss": 0.21420815, + "memory(GiB)": 45.23, + "step": 400, + "train_speed(iter/s)": 0.764075 + }, + { + "epoch": 0.03910614525139665, + "grad_norm": 1.34375, + "learning_rate": 5.21415270018622e-05, + "loss": 0.20427363, + "memory(GiB)": 45.23, + "step": 420, + "train_speed(iter/s)": 0.765517 + }, + { + "epoch": 0.040968342644320296, + "grad_norm": 1.265625, + "learning_rate": 5.462445685909373e-05, + "loss": 0.20555434, + "memory(GiB)": 45.23, + "step": 440, + "train_speed(iter/s)": 0.765864 + }, + { + "epoch": 0.04283054003724395, + "grad_norm": 1.34375, + "learning_rate": 5.7107386716325263e-05, + "loss": 0.21655397, + "memory(GiB)": 45.23, + "step": 460, + "train_speed(iter/s)": 0.766872 + }, + { + "epoch": 0.0446927374301676, + "grad_norm": 1.6484375, + "learning_rate": 5.95903165735568e-05, + "loss": 0.19097025, + "memory(GiB)": 45.23, + "step": 480, + "train_speed(iter/s)": 0.768982 + }, + { + "epoch": 0.04655493482309125, + "grad_norm": 1.6796875, + "learning_rate": 6.207324643078833e-05, + "loss": 0.19996111, + "memory(GiB)": 45.23, + "step": 500, + "train_speed(iter/s)": 0.770034 + }, + { + "epoch": 0.048417132216014895, + "grad_norm": 1.484375, + "learning_rate": 6.455617628801986e-05, + "loss": 0.20735807, + "memory(GiB)": 45.23, + "step": 520, + "train_speed(iter/s)": 0.771256 + }, + { + "epoch": 0.05027932960893855, + "grad_norm": 1.8203125, + "learning_rate": 6.70391061452514e-05, + "loss": 0.20899239, + "memory(GiB)": 45.23, + "step": 540, + "train_speed(iter/s)": 0.773346 + }, + { + "epoch": 0.0521415270018622, + "grad_norm": 2.1875, + "learning_rate": 6.952203600248294e-05, + "loss": 0.21587, + "memory(GiB)": 45.23, + "step": 560, + "train_speed(iter/s)": 0.775101 + }, + { + "epoch": 0.054003724394785846, + "grad_norm": 1.6484375, + "learning_rate": 7.200496585971447e-05, + "loss": 0.20331898, + "memory(GiB)": 45.23, + "step": 580, + "train_speed(iter/s)": 0.776379 + }, + { + "epoch": 0.055865921787709494, + "grad_norm": 1.1796875, + "learning_rate": 7.4487895716946e-05, + "loss": 0.19293334, + "memory(GiB)": 45.23, + "step": 600, + "train_speed(iter/s)": 0.776493 + }, + { + "epoch": 0.05772811918063315, + "grad_norm": 1.8671875, + "learning_rate": 7.697082557417753e-05, + "loss": 0.23722024, + "memory(GiB)": 45.23, + "step": 620, + "train_speed(iter/s)": 0.777184 + }, + { + "epoch": 0.0595903165735568, + "grad_norm": 1.3984375, + "learning_rate": 7.945375543140908e-05, + "loss": 0.21692863, + "memory(GiB)": 45.23, + "step": 640, + "train_speed(iter/s)": 0.77775 + }, + { + "epoch": 0.061452513966480445, + "grad_norm": 1.6171875, + "learning_rate": 8.193668528864061e-05, + "loss": 0.22430882, + "memory(GiB)": 45.23, + "step": 660, + "train_speed(iter/s)": 0.778898 + }, + { + "epoch": 0.0633147113594041, + "grad_norm": 1.5546875, + "learning_rate": 8.441961514587213e-05, + "loss": 0.22223544, + "memory(GiB)": 45.23, + "step": 680, + "train_speed(iter/s)": 0.7801 + }, + { + "epoch": 0.06517690875232775, + "grad_norm": 1.6015625, + "learning_rate": 8.690254500310366e-05, + "loss": 0.23334203, + "memory(GiB)": 45.23, + "step": 700, + "train_speed(iter/s)": 0.781002 + }, + { + "epoch": 0.0670391061452514, + "grad_norm": 1.7890625, + "learning_rate": 8.938547486033519e-05, + "loss": 0.20012813, + "memory(GiB)": 45.23, + "step": 720, + "train_speed(iter/s)": 0.781167 + }, + { + "epoch": 0.06890130353817504, + "grad_norm": 2.0625, + "learning_rate": 9.186840471756673e-05, + "loss": 0.22085505, + "memory(GiB)": 45.23, + "step": 740, + "train_speed(iter/s)": 0.780622 + }, + { + "epoch": 0.07076350093109869, + "grad_norm": 1.4140625, + "learning_rate": 9.435133457479826e-05, + "loss": 0.213342, + "memory(GiB)": 45.23, + "step": 760, + "train_speed(iter/s)": 0.781779 + }, + { + "epoch": 0.07262569832402235, + "grad_norm": 1.4921875, + "learning_rate": 9.68342644320298e-05, + "loss": 0.21715076, + "memory(GiB)": 45.23, + "step": 780, + "train_speed(iter/s)": 0.782338 + }, + { + "epoch": 0.074487895716946, + "grad_norm": 1.671875, + "learning_rate": 9.931719428926133e-05, + "loss": 0.22635069, + "memory(GiB)": 45.23, + "step": 800, + "train_speed(iter/s)": 0.782662 + }, + { + "epoch": 0.07635009310986965, + "grad_norm": 1.5234375, + "learning_rate": 0.00010180012414649287, + "loss": 0.20717359, + "memory(GiB)": 45.23, + "step": 820, + "train_speed(iter/s)": 0.783065 + }, + { + "epoch": 0.0782122905027933, + "grad_norm": 1.7890625, + "learning_rate": 0.0001042830540037244, + "loss": 0.22582912, + "memory(GiB)": 45.23, + "step": 840, + "train_speed(iter/s)": 0.783863 + }, + { + "epoch": 0.08007448789571694, + "grad_norm": 1.7578125, + "learning_rate": 0.00010676598386095593, + "loss": 0.22400272, + "memory(GiB)": 45.23, + "step": 860, + "train_speed(iter/s)": 0.784412 + }, + { + "epoch": 0.08193668528864059, + "grad_norm": 1.28125, + "learning_rate": 0.00010924891371818746, + "loss": 0.23082299, + "memory(GiB)": 45.23, + "step": 880, + "train_speed(iter/s)": 0.784183 + }, + { + "epoch": 0.08379888268156424, + "grad_norm": 1.46875, + "learning_rate": 0.000111731843575419, + "loss": 0.23772116, + "memory(GiB)": 45.23, + "step": 900, + "train_speed(iter/s)": 0.784761 + }, + { + "epoch": 0.0856610800744879, + "grad_norm": 1.796875, + "learning_rate": 0.00011421477343265053, + "loss": 0.23279626, + "memory(GiB)": 45.23, + "step": 920, + "train_speed(iter/s)": 0.785722 + }, + { + "epoch": 0.08752327746741155, + "grad_norm": 1.59375, + "learning_rate": 0.00011669770328988207, + "loss": 0.22019596, + "memory(GiB)": 45.23, + "step": 940, + "train_speed(iter/s)": 0.786441 + }, + { + "epoch": 0.0893854748603352, + "grad_norm": 1.6171875, + "learning_rate": 0.0001191806331471136, + "loss": 0.22586164, + "memory(GiB)": 45.23, + "step": 960, + "train_speed(iter/s)": 0.787059 + }, + { + "epoch": 0.09124767225325885, + "grad_norm": 1.7265625, + "learning_rate": 0.00012166356300434513, + "loss": 0.23982925, + "memory(GiB)": 45.23, + "step": 980, + "train_speed(iter/s)": 0.786758 + }, + { + "epoch": 0.0931098696461825, + "grad_norm": 2.234375, + "learning_rate": 0.00012414649286157667, + "loss": 0.22995653, + "memory(GiB)": 45.23, + "step": 1000, + "train_speed(iter/s)": 0.787241 + }, + { + "epoch": 0.09497206703910614, + "grad_norm": 1.8125, + "learning_rate": 0.0001266294227188082, + "loss": 0.25532598, + "memory(GiB)": 45.23, + "step": 1020, + "train_speed(iter/s)": 0.788014 + }, + { + "epoch": 0.09683426443202979, + "grad_norm": 1.5703125, + "learning_rate": 0.00012911235257603973, + "loss": 0.22124271, + "memory(GiB)": 45.23, + "step": 1040, + "train_speed(iter/s)": 0.788227 + }, + { + "epoch": 0.09869646182495345, + "grad_norm": 1.578125, + "learning_rate": 0.00013159528243327127, + "loss": 0.22865233, + "memory(GiB)": 45.23, + "step": 1060, + "train_speed(iter/s)": 0.788926 + }, + { + "epoch": 0.1005586592178771, + "grad_norm": 1.71875, + "learning_rate": 0.0001340782122905028, + "loss": 0.25925579, + "memory(GiB)": 45.23, + "step": 1080, + "train_speed(iter/s)": 0.788974 + }, + { + "epoch": 0.10242085661080075, + "grad_norm": 1.71875, + "learning_rate": 0.00013656114214773433, + "loss": 0.24655545, + "memory(GiB)": 45.23, + "step": 1100, + "train_speed(iter/s)": 0.789406 + }, + { + "epoch": 0.1042830540037244, + "grad_norm": 1.4921875, + "learning_rate": 0.00013904407200496588, + "loss": 0.25059803, + "memory(GiB)": 45.23, + "step": 1120, + "train_speed(iter/s)": 0.789646 + }, + { + "epoch": 0.10614525139664804, + "grad_norm": 1.90625, + "learning_rate": 0.0001415270018621974, + "loss": 0.26447239, + "memory(GiB)": 45.23, + "step": 1140, + "train_speed(iter/s)": 0.790032 + }, + { + "epoch": 0.10800744878957169, + "grad_norm": 1.40625, + "learning_rate": 0.00014400993171942894, + "loss": 0.28571393, + "memory(GiB)": 45.23, + "step": 1160, + "train_speed(iter/s)": 0.790468 + }, + { + "epoch": 0.10986964618249534, + "grad_norm": 2.140625, + "learning_rate": 0.00014649286157666046, + "loss": 0.26289692, + "memory(GiB)": 45.23, + "step": 1180, + "train_speed(iter/s)": 0.790894 + }, + { + "epoch": 0.11173184357541899, + "grad_norm": 1.6875, + "learning_rate": 0.000148975791433892, + "loss": 0.26204472, + "memory(GiB)": 45.23, + "step": 1200, + "train_speed(iter/s)": 0.791024 + }, + { + "epoch": 0.11359404096834265, + "grad_norm": 1.84375, + "learning_rate": 0.00015145872129112352, + "loss": 0.29556036, + "memory(GiB)": 45.23, + "step": 1220, + "train_speed(iter/s)": 0.791423 + }, + { + "epoch": 0.1154562383612663, + "grad_norm": 1.6328125, + "learning_rate": 0.00015394165114835507, + "loss": 0.27372065, + "memory(GiB)": 45.23, + "step": 1240, + "train_speed(iter/s)": 0.79099 + }, + { + "epoch": 0.11731843575418995, + "grad_norm": 2.03125, + "learning_rate": 0.00015642458100558658, + "loss": 0.28276806, + "memory(GiB)": 45.23, + "step": 1260, + "train_speed(iter/s)": 0.791011 + }, + { + "epoch": 0.1191806331471136, + "grad_norm": 1.53125, + "learning_rate": 0.00015890751086281816, + "loss": 0.28502092, + "memory(GiB)": 45.23, + "step": 1280, + "train_speed(iter/s)": 0.791642 + }, + { + "epoch": 0.12104283054003724, + "grad_norm": 1.8046875, + "learning_rate": 0.00016139044072004967, + "loss": 0.29006686, + "memory(GiB)": 45.23, + "step": 1300, + "train_speed(iter/s)": 0.791155 + }, + { + "epoch": 0.12290502793296089, + "grad_norm": 1.9296875, + "learning_rate": 0.00016387337057728122, + "loss": 0.29351676, + "memory(GiB)": 45.23, + "step": 1320, + "train_speed(iter/s)": 0.79116 + }, + { + "epoch": 0.12476722532588454, + "grad_norm": 1.5546875, + "learning_rate": 0.00016635630043451273, + "loss": 0.30054893, + "memory(GiB)": 45.23, + "step": 1340, + "train_speed(iter/s)": 0.791442 + }, + { + "epoch": 0.1266294227188082, + "grad_norm": 1.6796875, + "learning_rate": 0.00016883923029174425, + "loss": 0.284747, + "memory(GiB)": 45.23, + "step": 1360, + "train_speed(iter/s)": 0.791642 + }, + { + "epoch": 0.12849162011173185, + "grad_norm": 1.84375, + "learning_rate": 0.0001713221601489758, + "loss": 0.28408041, + "memory(GiB)": 45.23, + "step": 1380, + "train_speed(iter/s)": 0.792049 + }, + { + "epoch": 0.1303538175046555, + "grad_norm": 2.1875, + "learning_rate": 0.00017380509000620731, + "loss": 0.30376158, + "memory(GiB)": 45.23, + "step": 1400, + "train_speed(iter/s)": 0.79165 + }, + { + "epoch": 0.13221601489757914, + "grad_norm": 1.6640625, + "learning_rate": 0.00017628801986343886, + "loss": 0.30356815, + "memory(GiB)": 45.23, + "step": 1420, + "train_speed(iter/s)": 0.791951 + }, + { + "epoch": 0.1340782122905028, + "grad_norm": 1.796875, + "learning_rate": 0.00017877094972067038, + "loss": 0.29253561, + "memory(GiB)": 45.23, + "step": 1440, + "train_speed(iter/s)": 0.792277 + }, + { + "epoch": 0.13594040968342644, + "grad_norm": 1.828125, + "learning_rate": 0.00018125387957790195, + "loss": 0.31560664, + "memory(GiB)": 45.23, + "step": 1460, + "train_speed(iter/s)": 0.792471 + }, + { + "epoch": 0.1378026070763501, + "grad_norm": 1.9609375, + "learning_rate": 0.00018373680943513347, + "loss": 0.34103947, + "memory(GiB)": 45.23, + "step": 1480, + "train_speed(iter/s)": 0.792883 + }, + { + "epoch": 0.13966480446927373, + "grad_norm": 1.8125, + "learning_rate": 0.000186219739292365, + "loss": 0.32740824, + "memory(GiB)": 45.23, + "step": 1500, + "train_speed(iter/s)": 0.793276 + }, + { + "epoch": 0.14152700186219738, + "grad_norm": 2.046875, + "learning_rate": 0.00018870266914959653, + "loss": 0.32411356, + "memory(GiB)": 45.23, + "step": 1520, + "train_speed(iter/s)": 0.793438 + }, + { + "epoch": 0.14338919925512103, + "grad_norm": 2.375, + "learning_rate": 0.00019118559900682807, + "loss": 0.33667076, + "memory(GiB)": 45.23, + "step": 1540, + "train_speed(iter/s)": 0.793845 + }, + { + "epoch": 0.1452513966480447, + "grad_norm": 1.7734375, + "learning_rate": 0.0001936685288640596, + "loss": 0.3620486, + "memory(GiB)": 45.23, + "step": 1560, + "train_speed(iter/s)": 0.793711 + }, + { + "epoch": 0.14711359404096835, + "grad_norm": 2.0, + "learning_rate": 0.00019615145872129114, + "loss": 0.35051939, + "memory(GiB)": 45.23, + "step": 1580, + "train_speed(iter/s)": 0.793904 + }, + { + "epoch": 0.148975791433892, + "grad_norm": 2.09375, + "learning_rate": 0.00019863438857852265, + "loss": 0.32083566, + "memory(GiB)": 45.23, + "step": 1600, + "train_speed(iter/s)": 0.794229 + }, + { + "epoch": 0.15083798882681565, + "grad_norm": 1.796875, + "learning_rate": 0.00019999995733650257, + "loss": 0.34570508, + "memory(GiB)": 45.23, + "step": 1620, + "train_speed(iter/s)": 0.794576 + }, + { + "epoch": 0.1527001862197393, + "grad_norm": 1.8046875, + "learning_rate": 0.00019999955703731584, + "loss": 0.37476125, + "memory(GiB)": 45.23, + "step": 1640, + "train_speed(iter/s)": 0.794974 + }, + { + "epoch": 0.15456238361266295, + "grad_norm": 3.046875, + "learning_rate": 0.00019999873537223758, + "loss": 0.34583857, + "memory(GiB)": 45.23, + "step": 1660, + "train_speed(iter/s)": 0.79501 + }, + { + "epoch": 0.1564245810055866, + "grad_norm": 2.21875, + "learning_rate": 0.00019999749234473, + "loss": 0.36670647, + "memory(GiB)": 45.23, + "step": 1680, + "train_speed(iter/s)": 0.794806 + }, + { + "epoch": 0.15828677839851024, + "grad_norm": 2.140625, + "learning_rate": 0.00019999582796003082, + "loss": 0.35735064, + "memory(GiB)": 61.09, + "step": 1700, + "train_speed(iter/s)": 0.794864 + }, + { + "epoch": 0.1601489757914339, + "grad_norm": 1.9375, + "learning_rate": 0.00019999374222515323, + "loss": 0.37268331, + "memory(GiB)": 61.09, + "step": 1720, + "train_speed(iter/s)": 0.794839 + }, + { + "epoch": 0.16201117318435754, + "grad_norm": 1.734375, + "learning_rate": 0.00019999123514888583, + "loss": 0.35943956, + "memory(GiB)": 61.09, + "step": 1740, + "train_speed(iter/s)": 0.795031 + }, + { + "epoch": 0.16387337057728119, + "grad_norm": 1.765625, + "learning_rate": 0.00019998830674179265, + "loss": 0.36007235, + "memory(GiB)": 61.09, + "step": 1760, + "train_speed(iter/s)": 0.794945 + }, + { + "epoch": 0.16573556797020483, + "grad_norm": 1.6953125, + "learning_rate": 0.00019998495701621302, + "loss": 0.36344244, + "memory(GiB)": 61.09, + "step": 1780, + "train_speed(iter/s)": 0.794986 + }, + { + "epoch": 0.16759776536312848, + "grad_norm": 1.84375, + "learning_rate": 0.00019998118598626164, + "loss": 0.37410173, + "memory(GiB)": 61.09, + "step": 1800, + "train_speed(iter/s)": 0.794841 + }, + { + "epoch": 0.16945996275605213, + "grad_norm": 1.9140625, + "learning_rate": 0.0001999769936678284, + "loss": 0.36024194, + "memory(GiB)": 61.09, + "step": 1820, + "train_speed(iter/s)": 0.795138 + }, + { + "epoch": 0.1713221601489758, + "grad_norm": 1.9765625, + "learning_rate": 0.00019997238007857834, + "loss": 0.36868992, + "memory(GiB)": 61.09, + "step": 1840, + "train_speed(iter/s)": 0.795246 + }, + { + "epoch": 0.17318435754189945, + "grad_norm": 2.0, + "learning_rate": 0.0001999673452379517, + "loss": 0.39632101, + "memory(GiB)": 61.09, + "step": 1860, + "train_speed(iter/s)": 0.795603 + }, + { + "epoch": 0.1750465549348231, + "grad_norm": 2.53125, + "learning_rate": 0.00019996188916716366, + "loss": 0.38238664, + "memory(GiB)": 61.09, + "step": 1880, + "train_speed(iter/s)": 0.795803 + }, + { + "epoch": 0.17690875232774675, + "grad_norm": 1.671875, + "learning_rate": 0.00019995601188920432, + "loss": 0.39695168, + "memory(GiB)": 61.09, + "step": 1900, + "train_speed(iter/s)": 0.795572 + }, + { + "epoch": 0.1787709497206704, + "grad_norm": 2.28125, + "learning_rate": 0.00019994971342883865, + "loss": 0.37256937, + "memory(GiB)": 61.09, + "step": 1920, + "train_speed(iter/s)": 0.795652 + }, + { + "epoch": 0.18063314711359404, + "grad_norm": 2.34375, + "learning_rate": 0.00019994299381260634, + "loss": 0.38835607, + "memory(GiB)": 61.09, + "step": 1940, + "train_speed(iter/s)": 0.795347 + }, + { + "epoch": 0.1824953445065177, + "grad_norm": 1.7890625, + "learning_rate": 0.00019993585306882165, + "loss": 0.37124941, + "memory(GiB)": 61.09, + "step": 1960, + "train_speed(iter/s)": 0.795281 + }, + { + "epoch": 0.18435754189944134, + "grad_norm": 1.625, + "learning_rate": 0.00019992829122757343, + "loss": 0.40771437, + "memory(GiB)": 61.09, + "step": 1980, + "train_speed(iter/s)": 0.795167 + }, + { + "epoch": 0.186219739292365, + "grad_norm": 1.6484375, + "learning_rate": 0.0001999203083207248, + "loss": 0.40394459, + "memory(GiB)": 61.09, + "step": 2000, + "train_speed(iter/s)": 0.795156 + }, + { + "epoch": 0.186219739292365, + "eval_loss": 0.4044577181339264, + "eval_runtime": 77.5215, + "eval_samples_per_second": 179.112, + "eval_steps_per_second": 1.406, + "step": 2000 + }, + { + "epoch": 0.18808193668528864, + "grad_norm": 1.8125, + "learning_rate": 0.00019991190438191315, + "loss": 0.39065728, + "memory(GiB)": 61.09, + "step": 2020, + "train_speed(iter/s)": 0.762049 + }, + { + "epoch": 0.18994413407821228, + "grad_norm": 2.28125, + "learning_rate": 0.00019990307944654995, + "loss": 0.38545542, + "memory(GiB)": 61.09, + "step": 2040, + "train_speed(iter/s)": 0.762203 + }, + { + "epoch": 0.19180633147113593, + "grad_norm": 1.6796875, + "learning_rate": 0.0001998938335518207, + "loss": 0.39323997, + "memory(GiB)": 61.09, + "step": 2060, + "train_speed(iter/s)": 0.762592 + }, + { + "epoch": 0.19366852886405958, + "grad_norm": 1.8125, + "learning_rate": 0.0001998841667366846, + "loss": 0.40772767, + "memory(GiB)": 61.09, + "step": 2080, + "train_speed(iter/s)": 0.76303 + }, + { + "epoch": 0.19553072625698323, + "grad_norm": 1.5859375, + "learning_rate": 0.00019987407904187445, + "loss": 0.37801023, + "memory(GiB)": 61.09, + "step": 2100, + "train_speed(iter/s)": 0.763483 + }, + { + "epoch": 0.1973929236499069, + "grad_norm": 1.9140625, + "learning_rate": 0.00019986357050989656, + "loss": 0.37646284, + "memory(GiB)": 61.09, + "step": 2120, + "train_speed(iter/s)": 0.764105 + }, + { + "epoch": 0.19925512104283055, + "grad_norm": 2.390625, + "learning_rate": 0.0001998526411850305, + "loss": 0.39433451, + "memory(GiB)": 61.09, + "step": 2140, + "train_speed(iter/s)": 0.764481 + }, + { + "epoch": 0.2011173184357542, + "grad_norm": 2.1875, + "learning_rate": 0.00019984129111332895, + "loss": 0.37200718, + "memory(GiB)": 61.09, + "step": 2160, + "train_speed(iter/s)": 0.764802 + }, + { + "epoch": 0.20297951582867785, + "grad_norm": 2.28125, + "learning_rate": 0.00019982952034261744, + "loss": 0.39085593, + "memory(GiB)": 61.09, + "step": 2180, + "train_speed(iter/s)": 0.765212 + }, + { + "epoch": 0.2048417132216015, + "grad_norm": 2.046875, + "learning_rate": 0.00019981732892249414, + "loss": 0.41396513, + "memory(GiB)": 61.09, + "step": 2200, + "train_speed(iter/s)": 0.765596 + }, + { + "epoch": 0.20670391061452514, + "grad_norm": 1.859375, + "learning_rate": 0.00019980471690432986, + "loss": 0.38944612, + "memory(GiB)": 61.09, + "step": 2220, + "train_speed(iter/s)": 0.76564 + }, + { + "epoch": 0.2085661080074488, + "grad_norm": 1.515625, + "learning_rate": 0.00019979168434126752, + "loss": 0.3887634, + "memory(GiB)": 61.09, + "step": 2240, + "train_speed(iter/s)": 0.765877 + }, + { + "epoch": 0.21042830540037244, + "grad_norm": 2.265625, + "learning_rate": 0.00019977823128822215, + "loss": 0.40166087, + "memory(GiB)": 61.09, + "step": 2260, + "train_speed(iter/s)": 0.766381 + }, + { + "epoch": 0.2122905027932961, + "grad_norm": 2.171875, + "learning_rate": 0.00019976435780188058, + "loss": 0.39366639, + "memory(GiB)": 61.09, + "step": 2280, + "train_speed(iter/s)": 0.766888 + }, + { + "epoch": 0.21415270018621974, + "grad_norm": 2.078125, + "learning_rate": 0.00019975006394070118, + "loss": 0.3975667, + "memory(GiB)": 61.09, + "step": 2300, + "train_speed(iter/s)": 0.767231 + }, + { + "epoch": 0.21601489757914338, + "grad_norm": 1.6171875, + "learning_rate": 0.00019973534976491374, + "loss": 0.4033679, + "memory(GiB)": 61.09, + "step": 2320, + "train_speed(iter/s)": 0.767385 + }, + { + "epoch": 0.21787709497206703, + "grad_norm": 2.0, + "learning_rate": 0.000199720215336519, + "loss": 0.39846911, + "memory(GiB)": 61.09, + "step": 2340, + "train_speed(iter/s)": 0.767652 + }, + { + "epoch": 0.21973929236499068, + "grad_norm": 1.9453125, + "learning_rate": 0.0001997046607192886, + "loss": 0.38012104, + "memory(GiB)": 61.09, + "step": 2360, + "train_speed(iter/s)": 0.767841 + }, + { + "epoch": 0.22160148975791433, + "grad_norm": 2.265625, + "learning_rate": 0.0001996886859787646, + "loss": 0.40880508, + "memory(GiB)": 61.09, + "step": 2380, + "train_speed(iter/s)": 0.768088 + }, + { + "epoch": 0.22346368715083798, + "grad_norm": 2.484375, + "learning_rate": 0.0001996722911822595, + "loss": 0.40787139, + "memory(GiB)": 61.09, + "step": 2400, + "train_speed(iter/s)": 0.76849 + }, + { + "epoch": 0.22532588454376165, + "grad_norm": 1.640625, + "learning_rate": 0.00019965547639885564, + "loss": 0.37931633, + "memory(GiB)": 61.09, + "step": 2420, + "train_speed(iter/s)": 0.768893 + }, + { + "epoch": 0.2271880819366853, + "grad_norm": 2.0, + "learning_rate": 0.0001996382416994051, + "loss": 0.40943422, + "memory(GiB)": 61.09, + "step": 2440, + "train_speed(iter/s)": 0.76922 + }, + { + "epoch": 0.22905027932960895, + "grad_norm": 1.7578125, + "learning_rate": 0.00019962058715652936, + "loss": 0.39480171, + "memory(GiB)": 61.09, + "step": 2460, + "train_speed(iter/s)": 0.769501 + }, + { + "epoch": 0.2309124767225326, + "grad_norm": 1.84375, + "learning_rate": 0.00019960251284461896, + "loss": 0.39054556, + "memory(GiB)": 61.09, + "step": 2480, + "train_speed(iter/s)": 0.769688 + }, + { + "epoch": 0.23277467411545624, + "grad_norm": 2.3125, + "learning_rate": 0.00019958401883983325, + "loss": 0.41470366, + "memory(GiB)": 61.09, + "step": 2500, + "train_speed(iter/s)": 0.769923 + }, + { + "epoch": 0.2346368715083799, + "grad_norm": 2.40625, + "learning_rate": 0.00019956510522009993, + "loss": 0.40362306, + "memory(GiB)": 61.09, + "step": 2520, + "train_speed(iter/s)": 0.770155 + }, + { + "epoch": 0.23649906890130354, + "grad_norm": 1.6328125, + "learning_rate": 0.00019954577206511497, + "loss": 0.41036525, + "memory(GiB)": 61.09, + "step": 2540, + "train_speed(iter/s)": 0.770394 + }, + { + "epoch": 0.2383612662942272, + "grad_norm": 2.859375, + "learning_rate": 0.00019952601945634203, + "loss": 0.37247095, + "memory(GiB)": 61.09, + "step": 2560, + "train_speed(iter/s)": 0.770637 + }, + { + "epoch": 0.24022346368715083, + "grad_norm": 1.859375, + "learning_rate": 0.00019950584747701223, + "loss": 0.39711597, + "memory(GiB)": 61.09, + "step": 2580, + "train_speed(iter/s)": 0.770483 + }, + { + "epoch": 0.24208566108007448, + "grad_norm": 2.3125, + "learning_rate": 0.00019948525621212376, + "loss": 0.40421638, + "memory(GiB)": 61.09, + "step": 2600, + "train_speed(iter/s)": 0.770381 + }, + { + "epoch": 0.24394785847299813, + "grad_norm": 1.9296875, + "learning_rate": 0.00019946424574844164, + "loss": 0.38328767, + "memory(GiB)": 61.09, + "step": 2620, + "train_speed(iter/s)": 0.770361 + }, + { + "epoch": 0.24581005586592178, + "grad_norm": 1.65625, + "learning_rate": 0.00019944281617449713, + "loss": 0.40005746, + "memory(GiB)": 61.09, + "step": 2640, + "train_speed(iter/s)": 0.770354 + }, + { + "epoch": 0.24767225325884543, + "grad_norm": 2.453125, + "learning_rate": 0.00019942096758058758, + "loss": 0.43671055, + "memory(GiB)": 61.09, + "step": 2660, + "train_speed(iter/s)": 0.770409 + }, + { + "epoch": 0.24953445065176907, + "grad_norm": 2.09375, + "learning_rate": 0.00019939870005877589, + "loss": 0.40949144, + "memory(GiB)": 61.09, + "step": 2680, + "train_speed(iter/s)": 0.770232 + }, + { + "epoch": 0.25139664804469275, + "grad_norm": 1.609375, + "learning_rate": 0.00019937601370289022, + "loss": 0.39216471, + "memory(GiB)": 61.09, + "step": 2700, + "train_speed(iter/s)": 0.7704 + }, + { + "epoch": 0.2532588454376164, + "grad_norm": 2.015625, + "learning_rate": 0.0001993529086085236, + "loss": 0.42032986, + "memory(GiB)": 61.09, + "step": 2720, + "train_speed(iter/s)": 0.770556 + }, + { + "epoch": 0.25512104283054005, + "grad_norm": 1.6640625, + "learning_rate": 0.0001993293848730334, + "loss": 0.39470253, + "memory(GiB)": 61.09, + "step": 2740, + "train_speed(iter/s)": 0.770526 + }, + { + "epoch": 0.2569832402234637, + "grad_norm": 1.625, + "learning_rate": 0.00019930544259554107, + "loss": 0.40104327, + "memory(GiB)": 61.09, + "step": 2760, + "train_speed(iter/s)": 0.77006 + }, + { + "epoch": 0.25884543761638734, + "grad_norm": 1.4296875, + "learning_rate": 0.00019928108187693166, + "loss": 0.38365974, + "memory(GiB)": 61.09, + "step": 2780, + "train_speed(iter/s)": 0.769969 + }, + { + "epoch": 0.260707635009311, + "grad_norm": 2.46875, + "learning_rate": 0.0001992563028198534, + "loss": 0.41973991, + "memory(GiB)": 61.09, + "step": 2800, + "train_speed(iter/s)": 0.770041 + }, + { + "epoch": 0.26256983240223464, + "grad_norm": 1.59375, + "learning_rate": 0.0001992311055287172, + "loss": 0.39436083, + "memory(GiB)": 61.09, + "step": 2820, + "train_speed(iter/s)": 0.769916 + }, + { + "epoch": 0.2644320297951583, + "grad_norm": 1.84375, + "learning_rate": 0.00019920549010969633, + "loss": 0.40017633, + "memory(GiB)": 61.09, + "step": 2840, + "train_speed(iter/s)": 0.770067 + }, + { + "epoch": 0.26629422718808193, + "grad_norm": 1.65625, + "learning_rate": 0.00019917945667072596, + "loss": 0.41224184, + "memory(GiB)": 61.09, + "step": 2860, + "train_speed(iter/s)": 0.770021 + }, + { + "epoch": 0.2681564245810056, + "grad_norm": 2.171875, + "learning_rate": 0.00019915300532150255, + "loss": 0.39522376, + "memory(GiB)": 61.09, + "step": 2880, + "train_speed(iter/s)": 0.769932 + }, + { + "epoch": 0.27001862197392923, + "grad_norm": 1.7421875, + "learning_rate": 0.00019912613617348359, + "loss": 0.3813251, + "memory(GiB)": 61.09, + "step": 2900, + "train_speed(iter/s)": 0.770051 + }, + { + "epoch": 0.2718808193668529, + "grad_norm": 2.0625, + "learning_rate": 0.000199098849339887, + "loss": 0.40877409, + "memory(GiB)": 61.09, + "step": 2920, + "train_speed(iter/s)": 0.77 + }, + { + "epoch": 0.2737430167597765, + "grad_norm": 1.8515625, + "learning_rate": 0.00019907114493569067, + "loss": 0.3886503, + "memory(GiB)": 61.09, + "step": 2940, + "train_speed(iter/s)": 0.769917 + }, + { + "epoch": 0.2756052141527002, + "grad_norm": 1.4609375, + "learning_rate": 0.00019904302307763208, + "loss": 0.39552131, + "memory(GiB)": 61.09, + "step": 2960, + "train_speed(iter/s)": 0.769804 + }, + { + "epoch": 0.2774674115456238, + "grad_norm": 2.140625, + "learning_rate": 0.00019901448388420762, + "loss": 0.42156706, + "memory(GiB)": 61.09, + "step": 2980, + "train_speed(iter/s)": 0.769718 + }, + { + "epoch": 0.27932960893854747, + "grad_norm": 1.984375, + "learning_rate": 0.00019898552747567226, + "loss": 0.42330751, + "memory(GiB)": 61.09, + "step": 3000, + "train_speed(iter/s)": 0.769751 + }, + { + "epoch": 0.2811918063314711, + "grad_norm": 1.75, + "learning_rate": 0.00019895615397403903, + "loss": 0.40717449, + "memory(GiB)": 61.09, + "step": 3020, + "train_speed(iter/s)": 0.769716 + }, + { + "epoch": 0.28305400372439476, + "grad_norm": 1.8828125, + "learning_rate": 0.00019892636350307832, + "loss": 0.40095501, + "memory(GiB)": 61.09, + "step": 3040, + "train_speed(iter/s)": 0.769668 + }, + { + "epoch": 0.2849162011173184, + "grad_norm": 1.8359375, + "learning_rate": 0.0001988961561883176, + "loss": 0.40677986, + "memory(GiB)": 61.09, + "step": 3060, + "train_speed(iter/s)": 0.769698 + }, + { + "epoch": 0.28677839851024206, + "grad_norm": 1.96875, + "learning_rate": 0.00019886553215704074, + "loss": 0.43705454, + "memory(GiB)": 61.09, + "step": 3080, + "train_speed(iter/s)": 0.769667 + }, + { + "epoch": 0.2886405959031657, + "grad_norm": 2.015625, + "learning_rate": 0.00019883449153828756, + "loss": 0.42250085, + "memory(GiB)": 61.09, + "step": 3100, + "train_speed(iter/s)": 0.769504 + }, + { + "epoch": 0.2905027932960894, + "grad_norm": 2.015625, + "learning_rate": 0.0001988030344628532, + "loss": 0.4198226, + "memory(GiB)": 61.09, + "step": 3120, + "train_speed(iter/s)": 0.76958 + }, + { + "epoch": 0.29236499068901306, + "grad_norm": 2.4375, + "learning_rate": 0.00019877116106328764, + "loss": 0.39872463, + "memory(GiB)": 61.09, + "step": 3140, + "train_speed(iter/s)": 0.769603 + }, + { + "epoch": 0.2942271880819367, + "grad_norm": 1.421875, + "learning_rate": 0.0001987388714738951, + "loss": 0.42246108, + "memory(GiB)": 61.09, + "step": 3160, + "train_speed(iter/s)": 0.769508 + }, + { + "epoch": 0.29608938547486036, + "grad_norm": 1.8984375, + "learning_rate": 0.00019870616583073352, + "loss": 0.40915985, + "memory(GiB)": 61.09, + "step": 3180, + "train_speed(iter/s)": 0.769716 + }, + { + "epoch": 0.297951582867784, + "grad_norm": 1.5703125, + "learning_rate": 0.0001986730442716139, + "loss": 0.3973412, + "memory(GiB)": 61.09, + "step": 3200, + "train_speed(iter/s)": 0.769624 + }, + { + "epoch": 0.29981378026070765, + "grad_norm": 2.03125, + "learning_rate": 0.00019863950693609986, + "loss": 0.40871854, + "memory(GiB)": 61.09, + "step": 3220, + "train_speed(iter/s)": 0.769628 + }, + { + "epoch": 0.3016759776536313, + "grad_norm": 3.96875, + "learning_rate": 0.0001986055539655069, + "loss": 0.40967684, + "memory(GiB)": 61.09, + "step": 3240, + "train_speed(iter/s)": 0.769597 + }, + { + "epoch": 0.30353817504655495, + "grad_norm": 1.546875, + "learning_rate": 0.00019857118550290186, + "loss": 0.40765038, + "memory(GiB)": 61.09, + "step": 3260, + "train_speed(iter/s)": 0.769707 + }, + { + "epoch": 0.3054003724394786, + "grad_norm": 1.46875, + "learning_rate": 0.0001985364016931024, + "loss": 0.39425554, + "memory(GiB)": 61.09, + "step": 3280, + "train_speed(iter/s)": 0.769702 + }, + { + "epoch": 0.30726256983240224, + "grad_norm": 1.734375, + "learning_rate": 0.00019850120268267624, + "loss": 0.43635969, + "memory(GiB)": 61.09, + "step": 3300, + "train_speed(iter/s)": 0.769851 + }, + { + "epoch": 0.3091247672253259, + "grad_norm": 1.984375, + "learning_rate": 0.0001984655886199407, + "loss": 0.38309081, + "memory(GiB)": 61.09, + "step": 3320, + "train_speed(iter/s)": 0.7701 + }, + { + "epoch": 0.31098696461824954, + "grad_norm": 1.7578125, + "learning_rate": 0.00019842955965496196, + "loss": 0.39160743, + "memory(GiB)": 61.09, + "step": 3340, + "train_speed(iter/s)": 0.770116 + }, + { + "epoch": 0.3128491620111732, + "grad_norm": 1.703125, + "learning_rate": 0.00019839311593955442, + "loss": 0.3894845, + "memory(GiB)": 61.09, + "step": 3360, + "train_speed(iter/s)": 0.770262 + }, + { + "epoch": 0.31471135940409684, + "grad_norm": 1.921875, + "learning_rate": 0.00019835625762728017, + "loss": 0.40365009, + "memory(GiB)": 61.09, + "step": 3380, + "train_speed(iter/s)": 0.770237 + }, + { + "epoch": 0.3165735567970205, + "grad_norm": 1.9140625, + "learning_rate": 0.00019831898487344827, + "loss": 0.39198034, + "memory(GiB)": 61.09, + "step": 3400, + "train_speed(iter/s)": 0.770325 + }, + { + "epoch": 0.31843575418994413, + "grad_norm": 1.8125, + "learning_rate": 0.00019828129783511404, + "loss": 0.40903959, + "memory(GiB)": 61.09, + "step": 3420, + "train_speed(iter/s)": 0.770274 + }, + { + "epoch": 0.3202979515828678, + "grad_norm": 1.78125, + "learning_rate": 0.00019824319667107859, + "loss": 0.39606228, + "memory(GiB)": 61.09, + "step": 3440, + "train_speed(iter/s)": 0.770391 + }, + { + "epoch": 0.3221601489757914, + "grad_norm": 1.703125, + "learning_rate": 0.00019820468154188781, + "loss": 0.40229883, + "memory(GiB)": 61.09, + "step": 3460, + "train_speed(iter/s)": 0.770531 + }, + { + "epoch": 0.3240223463687151, + "grad_norm": 1.5546875, + "learning_rate": 0.00019816575260983221, + "loss": 0.42437396, + "memory(GiB)": 61.09, + "step": 3480, + "train_speed(iter/s)": 0.770378 + }, + { + "epoch": 0.3258845437616387, + "grad_norm": 1.890625, + "learning_rate": 0.00019812641003894563, + "loss": 0.38859506, + "memory(GiB)": 61.09, + "step": 3500, + "train_speed(iter/s)": 0.770398 + }, + { + "epoch": 0.32774674115456237, + "grad_norm": 1.96875, + "learning_rate": 0.000198086653995005, + "loss": 0.40636191, + "memory(GiB)": 61.09, + "step": 3520, + "train_speed(iter/s)": 0.77057 + }, + { + "epoch": 0.329608938547486, + "grad_norm": 1.671875, + "learning_rate": 0.00019804648464552947, + "loss": 0.3950357, + "memory(GiB)": 61.09, + "step": 3540, + "train_speed(iter/s)": 0.770758 + }, + { + "epoch": 0.33147113594040967, + "grad_norm": 2.25, + "learning_rate": 0.00019800590215977978, + "loss": 0.39195304, + "memory(GiB)": 61.09, + "step": 3560, + "train_speed(iter/s)": 0.770732 + }, + { + "epoch": 0.3333333333333333, + "grad_norm": 2.125, + "learning_rate": 0.0001979649067087574, + "loss": 0.39345875, + "memory(GiB)": 61.09, + "step": 3580, + "train_speed(iter/s)": 0.770863 + }, + { + "epoch": 0.33519553072625696, + "grad_norm": 2.21875, + "learning_rate": 0.00019792349846520395, + "loss": 0.42576008, + "memory(GiB)": 61.09, + "step": 3600, + "train_speed(iter/s)": 0.770989 + }, + { + "epoch": 0.3370577281191806, + "grad_norm": 1.6875, + "learning_rate": 0.00019788167760360038, + "loss": 0.39812212, + "memory(GiB)": 61.09, + "step": 3620, + "train_speed(iter/s)": 0.770888 + }, + { + "epoch": 0.33891992551210426, + "grad_norm": 1.6015625, + "learning_rate": 0.00019783944430016637, + "loss": 0.39472914, + "memory(GiB)": 61.09, + "step": 3640, + "train_speed(iter/s)": 0.77104 + }, + { + "epoch": 0.3407821229050279, + "grad_norm": 1.5703125, + "learning_rate": 0.00019779679873285942, + "loss": 0.37525396, + "memory(GiB)": 61.09, + "step": 3660, + "train_speed(iter/s)": 0.771096 + }, + { + "epoch": 0.3426443202979516, + "grad_norm": 2.40625, + "learning_rate": 0.00019775374108137422, + "loss": 0.38320711, + "memory(GiB)": 61.09, + "step": 3680, + "train_speed(iter/s)": 0.771183 + }, + { + "epoch": 0.34450651769087526, + "grad_norm": 1.96875, + "learning_rate": 0.0001977102715271418, + "loss": 0.43915482, + "memory(GiB)": 61.09, + "step": 3700, + "train_speed(iter/s)": 0.771154 + }, + { + "epoch": 0.3463687150837989, + "grad_norm": 2.0625, + "learning_rate": 0.00019766639025332888, + "loss": 0.38033762, + "memory(GiB)": 61.09, + "step": 3720, + "train_speed(iter/s)": 0.771208 + }, + { + "epoch": 0.34823091247672255, + "grad_norm": 1.6640625, + "learning_rate": 0.00019762209744483698, + "loss": 0.39966769, + "memory(GiB)": 61.09, + "step": 3740, + "train_speed(iter/s)": 0.771161 + }, + { + "epoch": 0.3500931098696462, + "grad_norm": 1.8046875, + "learning_rate": 0.0001975773932883017, + "loss": 0.40170197, + "memory(GiB)": 61.09, + "step": 3760, + "train_speed(iter/s)": 0.77126 + }, + { + "epoch": 0.35195530726256985, + "grad_norm": 1.7265625, + "learning_rate": 0.00019753227797209198, + "loss": 0.38694088, + "memory(GiB)": 61.09, + "step": 3780, + "train_speed(iter/s)": 0.771333 + }, + { + "epoch": 0.3538175046554935, + "grad_norm": 2.71875, + "learning_rate": 0.00019748675168630917, + "loss": 0.41248221, + "memory(GiB)": 61.09, + "step": 3800, + "train_speed(iter/s)": 0.771472 + }, + { + "epoch": 0.35567970204841715, + "grad_norm": 3.296875, + "learning_rate": 0.0001974408146227864, + "loss": 0.39216695, + "memory(GiB)": 61.09, + "step": 3820, + "train_speed(iter/s)": 0.771659 + }, + { + "epoch": 0.3575418994413408, + "grad_norm": 1.2890625, + "learning_rate": 0.0001973944669750876, + "loss": 0.39854093, + "memory(GiB)": 61.09, + "step": 3840, + "train_speed(iter/s)": 0.771931 + }, + { + "epoch": 0.35940409683426444, + "grad_norm": 1.671875, + "learning_rate": 0.00019734770893850686, + "loss": 0.40031991, + "memory(GiB)": 61.09, + "step": 3860, + "train_speed(iter/s)": 0.772166 + }, + { + "epoch": 0.3612662942271881, + "grad_norm": 1.9375, + "learning_rate": 0.00019730054071006748, + "loss": 0.4241601, + "memory(GiB)": 61.09, + "step": 3880, + "train_speed(iter/s)": 0.772192 + }, + { + "epoch": 0.36312849162011174, + "grad_norm": 1.7109375, + "learning_rate": 0.0001972529624885211, + "loss": 0.38440259, + "memory(GiB)": 61.09, + "step": 3900, + "train_speed(iter/s)": 0.772255 + }, + { + "epoch": 0.3649906890130354, + "grad_norm": 1.84375, + "learning_rate": 0.00019720497447434706, + "loss": 0.4149951, + "memory(GiB)": 61.09, + "step": 3920, + "train_speed(iter/s)": 0.77238 + }, + { + "epoch": 0.36685288640595903, + "grad_norm": 1.953125, + "learning_rate": 0.00019715657686975138, + "loss": 0.37301335, + "memory(GiB)": 61.09, + "step": 3940, + "train_speed(iter/s)": 0.772593 + }, + { + "epoch": 0.3687150837988827, + "grad_norm": 1.953125, + "learning_rate": 0.00019710776987866594, + "loss": 0.40297618, + "memory(GiB)": 61.09, + "step": 3960, + "train_speed(iter/s)": 0.772734 + }, + { + "epoch": 0.37057728119180633, + "grad_norm": 1.9375, + "learning_rate": 0.00019705855370674767, + "loss": 0.39141912, + "memory(GiB)": 61.09, + "step": 3980, + "train_speed(iter/s)": 0.772939 + }, + { + "epoch": 0.37243947858473, + "grad_norm": 1.96875, + "learning_rate": 0.00019700892856137765, + "loss": 0.39458523, + "memory(GiB)": 61.09, + "step": 4000, + "train_speed(iter/s)": 0.773013 + }, + { + "epoch": 0.37243947858473, + "eval_loss": 0.43771594762802124, + "eval_runtime": 76.1164, + "eval_samples_per_second": 182.418, + "eval_steps_per_second": 1.432, + "step": 4000 + }, + { + "epoch": 0.3743016759776536, + "grad_norm": 1.9140625, + "learning_rate": 0.00019695889465166024, + "loss": 0.40729074, + "memory(GiB)": 61.09, + "step": 4020, + "train_speed(iter/s)": 0.757361 + }, + { + "epoch": 0.3761638733705773, + "grad_norm": 1.7890625, + "learning_rate": 0.0001969084521884222, + "loss": 0.40453234, + "memory(GiB)": 61.09, + "step": 4040, + "train_speed(iter/s)": 0.757463 + }, + { + "epoch": 0.3780260707635009, + "grad_norm": 1.9375, + "learning_rate": 0.00019685760138421181, + "loss": 0.37667882, + "memory(GiB)": 61.09, + "step": 4060, + "train_speed(iter/s)": 0.757733 + }, + { + "epoch": 0.37988826815642457, + "grad_norm": 2.09375, + "learning_rate": 0.00019680634245329798, + "loss": 0.38365791, + "memory(GiB)": 61.09, + "step": 4080, + "train_speed(iter/s)": 0.757943 + }, + { + "epoch": 0.3817504655493482, + "grad_norm": 1.7109375, + "learning_rate": 0.0001967546756116693, + "loss": 0.42101908, + "memory(GiB)": 61.09, + "step": 4100, + "train_speed(iter/s)": 0.758279 + }, + { + "epoch": 0.38361266294227186, + "grad_norm": 2.09375, + "learning_rate": 0.0001967026010770332, + "loss": 0.40190611, + "memory(GiB)": 61.09, + "step": 4120, + "train_speed(iter/s)": 0.758514 + }, + { + "epoch": 0.3854748603351955, + "grad_norm": 1.84375, + "learning_rate": 0.00019665011906881493, + "loss": 0.41981277, + "memory(GiB)": 61.09, + "step": 4140, + "train_speed(iter/s)": 0.75872 + }, + { + "epoch": 0.38733705772811916, + "grad_norm": 1.9765625, + "learning_rate": 0.0001965972298081568, + "loss": 0.38099627, + "memory(GiB)": 61.09, + "step": 4160, + "train_speed(iter/s)": 0.758977 + }, + { + "epoch": 0.3891992551210428, + "grad_norm": 1.875, + "learning_rate": 0.0001965439335179171, + "loss": 0.3816371, + "memory(GiB)": 61.09, + "step": 4180, + "train_speed(iter/s)": 0.759186 + }, + { + "epoch": 0.39106145251396646, + "grad_norm": 1.6796875, + "learning_rate": 0.00019649023042266924, + "loss": 0.39355574, + "memory(GiB)": 61.09, + "step": 4200, + "train_speed(iter/s)": 0.759226 + }, + { + "epoch": 0.3929236499068901, + "grad_norm": 1.7890625, + "learning_rate": 0.00019643612074870067, + "loss": 0.42333245, + "memory(GiB)": 61.09, + "step": 4220, + "train_speed(iter/s)": 0.75949 + }, + { + "epoch": 0.3947858472998138, + "grad_norm": 1.6015625, + "learning_rate": 0.00019638160472401216, + "loss": 0.38143654, + "memory(GiB)": 61.09, + "step": 4240, + "train_speed(iter/s)": 0.759673 + }, + { + "epoch": 0.39664804469273746, + "grad_norm": 1.7890625, + "learning_rate": 0.00019632668257831663, + "loss": 0.40620303, + "memory(GiB)": 61.09, + "step": 4260, + "train_speed(iter/s)": 0.759941 + }, + { + "epoch": 0.3985102420856611, + "grad_norm": 2.0625, + "learning_rate": 0.00019627135454303835, + "loss": 0.40267248, + "memory(GiB)": 61.09, + "step": 4280, + "train_speed(iter/s)": 0.760262 + }, + { + "epoch": 0.40037243947858475, + "grad_norm": 1.8203125, + "learning_rate": 0.00019621562085131179, + "loss": 0.40306973, + "memory(GiB)": 61.09, + "step": 4300, + "train_speed(iter/s)": 0.760521 + }, + { + "epoch": 0.4022346368715084, + "grad_norm": 2.125, + "learning_rate": 0.00019615948173798073, + "loss": 0.38690736, + "memory(GiB)": 61.09, + "step": 4320, + "train_speed(iter/s)": 0.76073 + }, + { + "epoch": 0.40409683426443205, + "grad_norm": 1.8828125, + "learning_rate": 0.0001961029374395973, + "loss": 0.40119696, + "memory(GiB)": 61.09, + "step": 4340, + "train_speed(iter/s)": 0.760992 + }, + { + "epoch": 0.4059590316573557, + "grad_norm": 1.8203125, + "learning_rate": 0.00019604598819442098, + "loss": 0.3982002, + "memory(GiB)": 61.09, + "step": 4360, + "train_speed(iter/s)": 0.761273 + }, + { + "epoch": 0.40782122905027934, + "grad_norm": 1.484375, + "learning_rate": 0.00019598863424241753, + "loss": 0.36936131, + "memory(GiB)": 61.09, + "step": 4380, + "train_speed(iter/s)": 0.76156 + }, + { + "epoch": 0.409683426443203, + "grad_norm": 2.015625, + "learning_rate": 0.000195930875825258, + "loss": 0.37518866, + "memory(GiB)": 61.09, + "step": 4400, + "train_speed(iter/s)": 0.76183 + }, + { + "epoch": 0.41154562383612664, + "grad_norm": 1.3828125, + "learning_rate": 0.0001958727131863177, + "loss": 0.39472764, + "memory(GiB)": 61.09, + "step": 4420, + "train_speed(iter/s)": 0.762004 + }, + { + "epoch": 0.4134078212290503, + "grad_norm": 1.6015625, + "learning_rate": 0.00019581414657067527, + "loss": 0.38476, + "memory(GiB)": 61.09, + "step": 4440, + "train_speed(iter/s)": 0.76225 + }, + { + "epoch": 0.41527001862197394, + "grad_norm": 2.03125, + "learning_rate": 0.00019575517622511153, + "loss": 0.44369678, + "memory(GiB)": 61.09, + "step": 4460, + "train_speed(iter/s)": 0.762464 + }, + { + "epoch": 0.4171322160148976, + "grad_norm": 1.8046875, + "learning_rate": 0.00019569580239810853, + "loss": 0.39095647, + "memory(GiB)": 61.09, + "step": 4480, + "train_speed(iter/s)": 0.762511 + }, + { + "epoch": 0.41899441340782123, + "grad_norm": 1.890625, + "learning_rate": 0.00019563602533984842, + "loss": 0.39244399, + "memory(GiB)": 61.09, + "step": 4500, + "train_speed(iter/s)": 0.762601 + }, + { + "epoch": 0.4208566108007449, + "grad_norm": 2.171875, + "learning_rate": 0.00019557584530221245, + "loss": 0.3898438, + "memory(GiB)": 61.09, + "step": 4520, + "train_speed(iter/s)": 0.762751 + }, + { + "epoch": 0.4227188081936685, + "grad_norm": 1.546875, + "learning_rate": 0.0001955152625387799, + "loss": 0.38730087, + "memory(GiB)": 61.09, + "step": 4540, + "train_speed(iter/s)": 0.762907 + }, + { + "epoch": 0.4245810055865922, + "grad_norm": 2.15625, + "learning_rate": 0.00019545427730482696, + "loss": 0.4196455, + "memory(GiB)": 61.09, + "step": 4560, + "train_speed(iter/s)": 0.763065 + }, + { + "epoch": 0.4264432029795158, + "grad_norm": 1.90625, + "learning_rate": 0.0001953928898573258, + "loss": 0.36989145, + "memory(GiB)": 61.09, + "step": 4580, + "train_speed(iter/s)": 0.763243 + }, + { + "epoch": 0.42830540037243947, + "grad_norm": 1.78125, + "learning_rate": 0.0001953311004549433, + "loss": 0.38576865, + "memory(GiB)": 61.09, + "step": 4600, + "train_speed(iter/s)": 0.763397 + }, + { + "epoch": 0.4301675977653631, + "grad_norm": 1.6796875, + "learning_rate": 0.0001952689093580401, + "loss": 0.37395325, + "memory(GiB)": 61.09, + "step": 4620, + "train_speed(iter/s)": 0.763554 + }, + { + "epoch": 0.43202979515828677, + "grad_norm": 1.7734375, + "learning_rate": 0.00019520631682866943, + "loss": 0.35659883, + "memory(GiB)": 61.09, + "step": 4640, + "train_speed(iter/s)": 0.763689 + }, + { + "epoch": 0.4338919925512104, + "grad_norm": 2.015625, + "learning_rate": 0.000195143323130576, + "loss": 0.37833457, + "memory(GiB)": 61.09, + "step": 4660, + "train_speed(iter/s)": 0.763955 + }, + { + "epoch": 0.43575418994413406, + "grad_norm": 1.3984375, + "learning_rate": 0.00019507992852919496, + "loss": 0.38467407, + "memory(GiB)": 61.09, + "step": 4680, + "train_speed(iter/s)": 0.76422 + }, + { + "epoch": 0.4376163873370577, + "grad_norm": 1.953125, + "learning_rate": 0.00019501613329165072, + "loss": 0.42341266, + "memory(GiB)": 61.09, + "step": 4700, + "train_speed(iter/s)": 0.764381 + }, + { + "epoch": 0.43947858472998136, + "grad_norm": 1.9140625, + "learning_rate": 0.00019495193768675586, + "loss": 0.39320512, + "memory(GiB)": 61.09, + "step": 4720, + "train_speed(iter/s)": 0.76442 + }, + { + "epoch": 0.441340782122905, + "grad_norm": 2.140625, + "learning_rate": 0.00019488734198500992, + "loss": 0.39242547, + "memory(GiB)": 61.09, + "step": 4740, + "train_speed(iter/s)": 0.764554 + }, + { + "epoch": 0.44320297951582865, + "grad_norm": 2.234375, + "learning_rate": 0.0001948223464585984, + "loss": 0.41076765, + "memory(GiB)": 61.09, + "step": 4760, + "train_speed(iter/s)": 0.764645 + }, + { + "epoch": 0.4450651769087523, + "grad_norm": 1.5234375, + "learning_rate": 0.0001947569513813915, + "loss": 0.39205194, + "memory(GiB)": 61.09, + "step": 4780, + "train_speed(iter/s)": 0.764725 + }, + { + "epoch": 0.44692737430167595, + "grad_norm": 1.5703125, + "learning_rate": 0.00019469115702894295, + "loss": 0.41007242, + "memory(GiB)": 61.09, + "step": 4800, + "train_speed(iter/s)": 0.764778 + }, + { + "epoch": 0.44878957169459965, + "grad_norm": 1.59375, + "learning_rate": 0.00019462496367848894, + "loss": 0.40207219, + "memory(GiB)": 61.09, + "step": 4820, + "train_speed(iter/s)": 0.764936 + }, + { + "epoch": 0.4506517690875233, + "grad_norm": 1.7109375, + "learning_rate": 0.00019455837160894693, + "loss": 0.37529342, + "memory(GiB)": 61.09, + "step": 4840, + "train_speed(iter/s)": 0.764978 + }, + { + "epoch": 0.45251396648044695, + "grad_norm": 1.7265625, + "learning_rate": 0.00019449138110091445, + "loss": 0.38617492, + "memory(GiB)": 61.09, + "step": 4860, + "train_speed(iter/s)": 0.765185 + }, + { + "epoch": 0.4543761638733706, + "grad_norm": 1.96875, + "learning_rate": 0.00019442399243666786, + "loss": 0.40544581, + "memory(GiB)": 61.09, + "step": 4880, + "train_speed(iter/s)": 0.765361 + }, + { + "epoch": 0.45623836126629425, + "grad_norm": 1.546875, + "learning_rate": 0.00019435620590016125, + "loss": 0.39625409, + "memory(GiB)": 61.09, + "step": 4900, + "train_speed(iter/s)": 0.765415 + }, + { + "epoch": 0.4581005586592179, + "grad_norm": 1.8984375, + "learning_rate": 0.00019428802177702529, + "loss": 0.37152231, + "memory(GiB)": 61.09, + "step": 4920, + "train_speed(iter/s)": 0.765621 + }, + { + "epoch": 0.45996275605214154, + "grad_norm": 2.171875, + "learning_rate": 0.00019421944035456586, + "loss": 0.40098886, + "memory(GiB)": 61.09, + "step": 4940, + "train_speed(iter/s)": 0.765762 + }, + { + "epoch": 0.4618249534450652, + "grad_norm": 1.4453125, + "learning_rate": 0.00019415046192176294, + "loss": 0.38226032, + "memory(GiB)": 61.09, + "step": 4960, + "train_speed(iter/s)": 0.765864 + }, + { + "epoch": 0.46368715083798884, + "grad_norm": 1.8046875, + "learning_rate": 0.00019408108676926942, + "loss": 0.4000948, + "memory(GiB)": 61.09, + "step": 4980, + "train_speed(iter/s)": 0.76605 + }, + { + "epoch": 0.4655493482309125, + "grad_norm": 1.625, + "learning_rate": 0.00019401131518940985, + "loss": 0.39587092, + "memory(GiB)": 61.09, + "step": 5000, + "train_speed(iter/s)": 0.766201 + }, + { + "epoch": 0.46741154562383613, + "grad_norm": 1.5234375, + "learning_rate": 0.00019394114747617916, + "loss": 0.37739358, + "memory(GiB)": 61.09, + "step": 5020, + "train_speed(iter/s)": 0.766333 + }, + { + "epoch": 0.4692737430167598, + "grad_norm": 1.6640625, + "learning_rate": 0.00019387058392524145, + "loss": 0.37761064, + "memory(GiB)": 61.09, + "step": 5040, + "train_speed(iter/s)": 0.766541 + }, + { + "epoch": 0.47113594040968343, + "grad_norm": 1.6484375, + "learning_rate": 0.0001937996248339288, + "loss": 0.36651943, + "memory(GiB)": 61.09, + "step": 5060, + "train_speed(iter/s)": 0.766602 + }, + { + "epoch": 0.4729981378026071, + "grad_norm": 1.9296875, + "learning_rate": 0.00019372827050123995, + "loss": 0.36461196, + "memory(GiB)": 61.09, + "step": 5080, + "train_speed(iter/s)": 0.766795 + }, + { + "epoch": 0.4748603351955307, + "grad_norm": 2.71875, + "learning_rate": 0.00019365652122783897, + "loss": 0.38799112, + "memory(GiB)": 61.09, + "step": 5100, + "train_speed(iter/s)": 0.766791 + }, + { + "epoch": 0.4767225325884544, + "grad_norm": 1.90625, + "learning_rate": 0.00019358437731605425, + "loss": 0.37337472, + "memory(GiB)": 61.09, + "step": 5120, + "train_speed(iter/s)": 0.766967 + }, + { + "epoch": 0.478584729981378, + "grad_norm": 1.5703125, + "learning_rate": 0.00019351183906987694, + "loss": 0.41348104, + "memory(GiB)": 61.09, + "step": 5140, + "train_speed(iter/s)": 0.767242 + }, + { + "epoch": 0.48044692737430167, + "grad_norm": 1.703125, + "learning_rate": 0.00019343890679495988, + "loss": 0.41450329, + "memory(GiB)": 61.09, + "step": 5160, + "train_speed(iter/s)": 0.767312 + }, + { + "epoch": 0.4823091247672253, + "grad_norm": 1.4921875, + "learning_rate": 0.00019336558079861607, + "loss": 0.35895565, + "memory(GiB)": 61.09, + "step": 5180, + "train_speed(iter/s)": 0.767456 + }, + { + "epoch": 0.48417132216014896, + "grad_norm": 1.875, + "learning_rate": 0.00019329186138981769, + "loss": 0.35140588, + "memory(GiB)": 61.09, + "step": 5200, + "train_speed(iter/s)": 0.767603 + }, + { + "epoch": 0.4860335195530726, + "grad_norm": 1.6484375, + "learning_rate": 0.0001932177488791945, + "loss": 0.39651599, + "memory(GiB)": 61.09, + "step": 5220, + "train_speed(iter/s)": 0.767676 + }, + { + "epoch": 0.48789571694599626, + "grad_norm": 1.6484375, + "learning_rate": 0.00019314324357903276, + "loss": 0.37009797, + "memory(GiB)": 61.09, + "step": 5240, + "train_speed(iter/s)": 0.767901 + }, + { + "epoch": 0.4897579143389199, + "grad_norm": 2.046875, + "learning_rate": 0.00019306834580327378, + "loss": 0.3629123, + "memory(GiB)": 61.09, + "step": 5260, + "train_speed(iter/s)": 0.768057 + }, + { + "epoch": 0.49162011173184356, + "grad_norm": 2.203125, + "learning_rate": 0.00019299305586751263, + "loss": 0.41036406, + "memory(GiB)": 61.09, + "step": 5280, + "train_speed(iter/s)": 0.768172 + }, + { + "epoch": 0.4934823091247672, + "grad_norm": 1.1953125, + "learning_rate": 0.0001929173740889968, + "loss": 0.3859787, + "memory(GiB)": 61.09, + "step": 5300, + "train_speed(iter/s)": 0.768303 + }, + { + "epoch": 0.49534450651769085, + "grad_norm": 1.421875, + "learning_rate": 0.00019284130078662498, + "loss": 0.36947384, + "memory(GiB)": 61.09, + "step": 5320, + "train_speed(iter/s)": 0.768419 + }, + { + "epoch": 0.4972067039106145, + "grad_norm": 1.765625, + "learning_rate": 0.00019276483628094548, + "loss": 0.39193997, + "memory(GiB)": 61.09, + "step": 5340, + "train_speed(iter/s)": 0.768551 + }, + { + "epoch": 0.49906890130353815, + "grad_norm": 1.703125, + "learning_rate": 0.00019268798089415508, + "loss": 0.40856562, + "memory(GiB)": 61.09, + "step": 5360, + "train_speed(iter/s)": 0.768667 + }, + { + "epoch": 0.5009310986964618, + "grad_norm": 1.75, + "learning_rate": 0.0001926107349500976, + "loss": 0.38102791, + "memory(GiB)": 61.09, + "step": 5380, + "train_speed(iter/s)": 0.768775 + }, + { + "epoch": 0.5027932960893855, + "grad_norm": 1.578125, + "learning_rate": 0.00019253309877426255, + "loss": 0.39010181, + "memory(GiB)": 61.09, + "step": 5400, + "train_speed(iter/s)": 0.768826 + }, + { + "epoch": 0.5046554934823091, + "grad_norm": 1.5625, + "learning_rate": 0.00019245507269378376, + "loss": 0.36907117, + "memory(GiB)": 61.09, + "step": 5420, + "train_speed(iter/s)": 0.768871 + }, + { + "epoch": 0.5065176908752328, + "grad_norm": 1.6875, + "learning_rate": 0.00019237665703743795, + "loss": 0.39245234, + "memory(GiB)": 61.09, + "step": 5440, + "train_speed(iter/s)": 0.769044 + }, + { + "epoch": 0.5083798882681564, + "grad_norm": 2.234375, + "learning_rate": 0.0001922978521356434, + "loss": 0.37116551, + "memory(GiB)": 61.09, + "step": 5460, + "train_speed(iter/s)": 0.769157 + }, + { + "epoch": 0.5102420856610801, + "grad_norm": 1.625, + "learning_rate": 0.0001922186583204586, + "loss": 0.3747494, + "memory(GiB)": 61.09, + "step": 5480, + "train_speed(iter/s)": 0.769282 + }, + { + "epoch": 0.5121042830540037, + "grad_norm": 1.421875, + "learning_rate": 0.00019213907592558068, + "loss": 0.39164233, + "memory(GiB)": 61.09, + "step": 5500, + "train_speed(iter/s)": 0.769399 + }, + { + "epoch": 0.5139664804469274, + "grad_norm": 1.734375, + "learning_rate": 0.0001920591052863443, + "loss": 0.36614444, + "memory(GiB)": 61.09, + "step": 5520, + "train_speed(iter/s)": 0.769497 + }, + { + "epoch": 0.515828677839851, + "grad_norm": 1.296875, + "learning_rate": 0.0001919787467397199, + "loss": 0.36284633, + "memory(GiB)": 61.09, + "step": 5540, + "train_speed(iter/s)": 0.769655 + }, + { + "epoch": 0.5176908752327747, + "grad_norm": 1.8984375, + "learning_rate": 0.00019189800062431248, + "loss": 0.38447697, + "memory(GiB)": 61.09, + "step": 5560, + "train_speed(iter/s)": 0.769671 + }, + { + "epoch": 0.5195530726256983, + "grad_norm": 1.6640625, + "learning_rate": 0.0001918168672803601, + "loss": 0.3774585, + "memory(GiB)": 61.09, + "step": 5580, + "train_speed(iter/s)": 0.769848 + }, + { + "epoch": 0.521415270018622, + "grad_norm": 1.6015625, + "learning_rate": 0.00019173534704973255, + "loss": 0.37147908, + "memory(GiB)": 61.09, + "step": 5600, + "train_speed(iter/s)": 0.769928 + }, + { + "epoch": 0.5232774674115456, + "grad_norm": 1.9140625, + "learning_rate": 0.00019165344027592985, + "loss": 0.4088347, + "memory(GiB)": 61.09, + "step": 5620, + "train_speed(iter/s)": 0.77009 + }, + { + "epoch": 0.5251396648044693, + "grad_norm": 1.796875, + "learning_rate": 0.00019157114730408064, + "loss": 0.38829794, + "memory(GiB)": 61.09, + "step": 5640, + "train_speed(iter/s)": 0.770242 + }, + { + "epoch": 0.527001862197393, + "grad_norm": 2.140625, + "learning_rate": 0.000191488468480941, + "loss": 0.36397715, + "memory(GiB)": 61.09, + "step": 5660, + "train_speed(iter/s)": 0.770341 + }, + { + "epoch": 0.5288640595903166, + "grad_norm": 1.421875, + "learning_rate": 0.0001914054041548929, + "loss": 0.38371773, + "memory(GiB)": 61.09, + "step": 5680, + "train_speed(iter/s)": 0.770391 + }, + { + "epoch": 0.5307262569832403, + "grad_norm": 1.2734375, + "learning_rate": 0.00019132195467594254, + "loss": 0.3827935, + "memory(GiB)": 61.09, + "step": 5700, + "train_speed(iter/s)": 0.77059 + }, + { + "epoch": 0.5325884543761639, + "grad_norm": 1.890625, + "learning_rate": 0.00019123812039571915, + "loss": 0.38209565, + "memory(GiB)": 61.09, + "step": 5720, + "train_speed(iter/s)": 0.770819 + }, + { + "epoch": 0.5344506517690876, + "grad_norm": 1.3828125, + "learning_rate": 0.0001911539016674734, + "loss": 0.35899367, + "memory(GiB)": 61.09, + "step": 5740, + "train_speed(iter/s)": 0.771004 + }, + { + "epoch": 0.5363128491620112, + "grad_norm": 1.625, + "learning_rate": 0.00019106929884607577, + "loss": 0.38742561, + "memory(GiB)": 61.09, + "step": 5760, + "train_speed(iter/s)": 0.770977 + }, + { + "epoch": 0.5381750465549349, + "grad_norm": 1.4609375, + "learning_rate": 0.00019098431228801535, + "loss": 0.40184608, + "memory(GiB)": 61.09, + "step": 5780, + "train_speed(iter/s)": 0.770971 + }, + { + "epoch": 0.5400372439478585, + "grad_norm": 1.5625, + "learning_rate": 0.00019089894235139806, + "loss": 0.37216854, + "memory(GiB)": 61.09, + "step": 5800, + "train_speed(iter/s)": 0.771075 + }, + { + "epoch": 0.5418994413407822, + "grad_norm": 1.625, + "learning_rate": 0.00019081318939594532, + "loss": 0.38967266, + "memory(GiB)": 61.09, + "step": 5820, + "train_speed(iter/s)": 0.77124 + }, + { + "epoch": 0.5437616387337058, + "grad_norm": 1.78125, + "learning_rate": 0.00019072705378299237, + "loss": 0.36322145, + "memory(GiB)": 61.09, + "step": 5840, + "train_speed(iter/s)": 0.771282 + }, + { + "epoch": 0.5456238361266295, + "grad_norm": 1.7734375, + "learning_rate": 0.00019064053587548697, + "loss": 0.41168041, + "memory(GiB)": 61.09, + "step": 5860, + "train_speed(iter/s)": 0.771409 + }, + { + "epoch": 0.547486033519553, + "grad_norm": 1.515625, + "learning_rate": 0.00019055363603798765, + "loss": 0.38845997, + "memory(GiB)": 61.09, + "step": 5880, + "train_speed(iter/s)": 0.771601 + }, + { + "epoch": 0.5493482309124768, + "grad_norm": 2.203125, + "learning_rate": 0.00019046635463666232, + "loss": 0.35851908, + "memory(GiB)": 61.09, + "step": 5900, + "train_speed(iter/s)": 0.771596 + }, + { + "epoch": 0.5512104283054003, + "grad_norm": 2.109375, + "learning_rate": 0.00019037869203928671, + "loss": 0.38949237, + "memory(GiB)": 61.09, + "step": 5920, + "train_speed(iter/s)": 0.771651 + }, + { + "epoch": 0.553072625698324, + "grad_norm": 1.4921875, + "learning_rate": 0.00019029064861524267, + "loss": 0.36254482, + "memory(GiB)": 61.09, + "step": 5940, + "train_speed(iter/s)": 0.771719 + }, + { + "epoch": 0.5549348230912476, + "grad_norm": 1.5859375, + "learning_rate": 0.00019020222473551686, + "loss": 0.37663541, + "memory(GiB)": 61.09, + "step": 5960, + "train_speed(iter/s)": 0.771885 + }, + { + "epoch": 0.5567970204841713, + "grad_norm": 1.796875, + "learning_rate": 0.00019011342077269894, + "loss": 0.37530179, + "memory(GiB)": 61.09, + "step": 5980, + "train_speed(iter/s)": 0.77194 + }, + { + "epoch": 0.5586592178770949, + "grad_norm": 1.8203125, + "learning_rate": 0.00019002423710098022, + "loss": 0.33494754, + "memory(GiB)": 61.09, + "step": 6000, + "train_speed(iter/s)": 0.771986 + }, + { + "epoch": 0.5586592178770949, + "eval_loss": 0.44819557666778564, + "eval_runtime": 76.4302, + "eval_samples_per_second": 181.669, + "eval_steps_per_second": 1.426, + "step": 6000 + }, + { + "epoch": 0.5605214152700186, + "grad_norm": 1.7109375, + "learning_rate": 0.00018993467409615192, + "loss": 0.39943161, + "memory(GiB)": 61.09, + "step": 6020, + "train_speed(iter/s)": 0.761403 + }, + { + "epoch": 0.5623836126629422, + "grad_norm": 1.5546875, + "learning_rate": 0.00018984473213560366, + "loss": 0.37351987, + "memory(GiB)": 61.09, + "step": 6040, + "train_speed(iter/s)": 0.76152 + }, + { + "epoch": 0.5642458100558659, + "grad_norm": 1.6484375, + "learning_rate": 0.0001897544115983219, + "loss": 0.39573755, + "memory(GiB)": 61.09, + "step": 6060, + "train_speed(iter/s)": 0.761669 + }, + { + "epoch": 0.5661080074487895, + "grad_norm": 1.703125, + "learning_rate": 0.00018966371286488823, + "loss": 0.4023766, + "memory(GiB)": 61.09, + "step": 6080, + "train_speed(iter/s)": 0.761822 + }, + { + "epoch": 0.5679702048417132, + "grad_norm": 2.15625, + "learning_rate": 0.00018957263631747787, + "loss": 0.3705966, + "memory(GiB)": 61.09, + "step": 6100, + "train_speed(iter/s)": 0.76184 + }, + { + "epoch": 0.5698324022346368, + "grad_norm": 2.25, + "learning_rate": 0.00018948118233985803, + "loss": 0.38215985, + "memory(GiB)": 61.09, + "step": 6120, + "train_speed(iter/s)": 0.76194 + }, + { + "epoch": 0.5716945996275605, + "grad_norm": 1.484375, + "learning_rate": 0.0001893893513173863, + "loss": 0.37131, + "memory(GiB)": 61.09, + "step": 6140, + "train_speed(iter/s)": 0.762158 + }, + { + "epoch": 0.5735567970204841, + "grad_norm": 1.5859375, + "learning_rate": 0.00018929714363700897, + "loss": 0.38114462, + "memory(GiB)": 61.09, + "step": 6160, + "train_speed(iter/s)": 0.762317 + }, + { + "epoch": 0.5754189944134078, + "grad_norm": 1.78125, + "learning_rate": 0.0001892045596872595, + "loss": 0.36428151, + "memory(GiB)": 61.09, + "step": 6180, + "train_speed(iter/s)": 0.762461 + }, + { + "epoch": 0.5772811918063314, + "grad_norm": 1.546875, + "learning_rate": 0.00018911159985825684, + "loss": 0.38127549, + "memory(GiB)": 61.09, + "step": 6200, + "train_speed(iter/s)": 0.762497 + }, + { + "epoch": 0.5791433891992551, + "grad_norm": 3.015625, + "learning_rate": 0.00018901826454170367, + "loss": 0.38451765, + "memory(GiB)": 61.09, + "step": 6220, + "train_speed(iter/s)": 0.762625 + }, + { + "epoch": 0.5810055865921788, + "grad_norm": 1.4609375, + "learning_rate": 0.00018892455413088493, + "loss": 0.39861171, + "memory(GiB)": 61.09, + "step": 6240, + "train_speed(iter/s)": 0.762782 + }, + { + "epoch": 0.5828677839851024, + "grad_norm": 1.6796875, + "learning_rate": 0.0001888304690206661, + "loss": 0.35783043, + "memory(GiB)": 61.09, + "step": 6260, + "train_speed(iter/s)": 0.762874 + }, + { + "epoch": 0.5847299813780261, + "grad_norm": 1.375, + "learning_rate": 0.00018873600960749148, + "loss": 0.36647081, + "memory(GiB)": 61.09, + "step": 6280, + "train_speed(iter/s)": 0.762938 + }, + { + "epoch": 0.5865921787709497, + "grad_norm": 1.7109375, + "learning_rate": 0.0001886411762893826, + "loss": 0.3527771, + "memory(GiB)": 61.09, + "step": 6300, + "train_speed(iter/s)": 0.763086 + }, + { + "epoch": 0.5884543761638734, + "grad_norm": 1.8515625, + "learning_rate": 0.0001885459694659365, + "loss": 0.38488383, + "memory(GiB)": 61.09, + "step": 6320, + "train_speed(iter/s)": 0.763164 + }, + { + "epoch": 0.590316573556797, + "grad_norm": 1.34375, + "learning_rate": 0.00018845038953832401, + "loss": 0.37899766, + "memory(GiB)": 61.09, + "step": 6340, + "train_speed(iter/s)": 0.763257 + }, + { + "epoch": 0.5921787709497207, + "grad_norm": 1.9296875, + "learning_rate": 0.00018835443690928813, + "loss": 0.36712158, + "memory(GiB)": 61.09, + "step": 6360, + "train_speed(iter/s)": 0.763341 + }, + { + "epoch": 0.5940409683426443, + "grad_norm": 1.8203125, + "learning_rate": 0.0001882581119831423, + "loss": 0.36925483, + "memory(GiB)": 61.09, + "step": 6380, + "train_speed(iter/s)": 0.7635 + }, + { + "epoch": 0.595903165735568, + "grad_norm": 1.421875, + "learning_rate": 0.00018816141516576874, + "loss": 0.37719622, + "memory(GiB)": 61.09, + "step": 6400, + "train_speed(iter/s)": 0.763631 + }, + { + "epoch": 0.5977653631284916, + "grad_norm": 1.9765625, + "learning_rate": 0.0001880643468646166, + "loss": 0.37256207, + "memory(GiB)": 61.09, + "step": 6420, + "train_speed(iter/s)": 0.76375 + }, + { + "epoch": 0.5996275605214153, + "grad_norm": 1.703125, + "learning_rate": 0.00018796690748870045, + "loss": 0.36438172, + "memory(GiB)": 61.09, + "step": 6440, + "train_speed(iter/s)": 0.763808 + }, + { + "epoch": 0.6014897579143389, + "grad_norm": 1.6328125, + "learning_rate": 0.00018786909744859838, + "loss": 0.37020493, + "memory(GiB)": 61.09, + "step": 6460, + "train_speed(iter/s)": 0.763949 + }, + { + "epoch": 0.6033519553072626, + "grad_norm": 1.4609375, + "learning_rate": 0.0001877709171564504, + "loss": 0.37200084, + "memory(GiB)": 61.09, + "step": 6480, + "train_speed(iter/s)": 0.763933 + }, + { + "epoch": 0.6052141527001862, + "grad_norm": 1.3359375, + "learning_rate": 0.00018767236702595653, + "loss": 0.37319822, + "memory(GiB)": 61.09, + "step": 6500, + "train_speed(iter/s)": 0.764126 + }, + { + "epoch": 0.6070763500931099, + "grad_norm": 1.6015625, + "learning_rate": 0.00018757344747237536, + "loss": 0.37853823, + "memory(GiB)": 61.09, + "step": 6520, + "train_speed(iter/s)": 0.764271 + }, + { + "epoch": 0.6089385474860335, + "grad_norm": 1.703125, + "learning_rate": 0.00018747415891252187, + "loss": 0.37537384, + "memory(GiB)": 61.09, + "step": 6540, + "train_speed(iter/s)": 0.764402 + }, + { + "epoch": 0.6108007448789572, + "grad_norm": 1.78125, + "learning_rate": 0.00018737450176476613, + "loss": 0.36597242, + "memory(GiB)": 61.09, + "step": 6560, + "train_speed(iter/s)": 0.764565 + }, + { + "epoch": 0.6126629422718808, + "grad_norm": 1.9375, + "learning_rate": 0.0001872744764490312, + "loss": 0.35904284, + "memory(GiB)": 61.09, + "step": 6580, + "train_speed(iter/s)": 0.764671 + }, + { + "epoch": 0.6145251396648045, + "grad_norm": 1.7734375, + "learning_rate": 0.00018717408338679147, + "loss": 0.3644063, + "memory(GiB)": 61.09, + "step": 6600, + "train_speed(iter/s)": 0.764735 + }, + { + "epoch": 0.6163873370577281, + "grad_norm": 1.953125, + "learning_rate": 0.00018707332300107102, + "loss": 0.35240235, + "memory(GiB)": 61.09, + "step": 6620, + "train_speed(iter/s)": 0.764831 + }, + { + "epoch": 0.6182495344506518, + "grad_norm": 1.59375, + "learning_rate": 0.00018697219571644155, + "loss": 0.36407018, + "memory(GiB)": 61.09, + "step": 6640, + "train_speed(iter/s)": 0.765007 + }, + { + "epoch": 0.6201117318435754, + "grad_norm": 1.4296875, + "learning_rate": 0.0001868707019590209, + "loss": 0.35534208, + "memory(GiB)": 61.09, + "step": 6660, + "train_speed(iter/s)": 0.765128 + }, + { + "epoch": 0.6219739292364991, + "grad_norm": 1.484375, + "learning_rate": 0.00018676884215647094, + "loss": 0.34723496, + "memory(GiB)": 61.09, + "step": 6680, + "train_speed(iter/s)": 0.76517 + }, + { + "epoch": 0.6238361266294227, + "grad_norm": 1.6875, + "learning_rate": 0.00018666661673799614, + "loss": 0.36451449, + "memory(GiB)": 61.09, + "step": 6700, + "train_speed(iter/s)": 0.765259 + }, + { + "epoch": 0.6256983240223464, + "grad_norm": 1.828125, + "learning_rate": 0.00018656402613434142, + "loss": 0.36834097, + "memory(GiB)": 61.09, + "step": 6720, + "train_speed(iter/s)": 0.765415 + }, + { + "epoch": 0.62756052141527, + "grad_norm": 1.8671875, + "learning_rate": 0.00018646107077779046, + "loss": 0.3686954, + "memory(GiB)": 61.09, + "step": 6740, + "train_speed(iter/s)": 0.765555 + }, + { + "epoch": 0.6294227188081937, + "grad_norm": 1.75, + "learning_rate": 0.00018635775110216403, + "loss": 0.35826502, + "memory(GiB)": 61.09, + "step": 6760, + "train_speed(iter/s)": 0.765672 + }, + { + "epoch": 0.6312849162011173, + "grad_norm": 1.8984375, + "learning_rate": 0.00018625406754281784, + "loss": 0.36378212, + "memory(GiB)": 61.09, + "step": 6780, + "train_speed(iter/s)": 0.765766 + }, + { + "epoch": 0.633147113594041, + "grad_norm": 1.8203125, + "learning_rate": 0.00018615002053664104, + "loss": 0.36196086, + "memory(GiB)": 61.09, + "step": 6800, + "train_speed(iter/s)": 0.765831 + }, + { + "epoch": 0.6350093109869647, + "grad_norm": 1.546875, + "learning_rate": 0.00018604561052205413, + "loss": 0.38385079, + "memory(GiB)": 61.09, + "step": 6820, + "train_speed(iter/s)": 0.765852 + }, + { + "epoch": 0.6368715083798883, + "grad_norm": 1.3984375, + "learning_rate": 0.0001859408379390073, + "loss": 0.34547548, + "memory(GiB)": 61.09, + "step": 6840, + "train_speed(iter/s)": 0.765936 + }, + { + "epoch": 0.638733705772812, + "grad_norm": 1.5546875, + "learning_rate": 0.00018583570322897836, + "loss": 0.37987022, + "memory(GiB)": 61.09, + "step": 6860, + "train_speed(iter/s)": 0.76609 + }, + { + "epoch": 0.6405959031657356, + "grad_norm": 1.375, + "learning_rate": 0.00018573020683497113, + "loss": 0.34121661, + "memory(GiB)": 61.09, + "step": 6880, + "train_speed(iter/s)": 0.766189 + }, + { + "epoch": 0.6424581005586593, + "grad_norm": 2.046875, + "learning_rate": 0.00018562434920151336, + "loss": 0.38715105, + "memory(GiB)": 61.09, + "step": 6900, + "train_speed(iter/s)": 0.766299 + }, + { + "epoch": 0.6443202979515829, + "grad_norm": 1.7265625, + "learning_rate": 0.00018551813077465504, + "loss": 0.36055377, + "memory(GiB)": 61.09, + "step": 6920, + "train_speed(iter/s)": 0.766405 + }, + { + "epoch": 0.6461824953445066, + "grad_norm": 1.984375, + "learning_rate": 0.00018541155200196637, + "loss": 0.36063504, + "memory(GiB)": 61.09, + "step": 6940, + "train_speed(iter/s)": 0.766488 + }, + { + "epoch": 0.6480446927374302, + "grad_norm": 1.5390625, + "learning_rate": 0.0001853046133325359, + "loss": 0.36382327, + "memory(GiB)": 61.09, + "step": 6960, + "train_speed(iter/s)": 0.766578 + }, + { + "epoch": 0.6499068901303539, + "grad_norm": 1.5703125, + "learning_rate": 0.00018519731521696872, + "loss": 0.37230091, + "memory(GiB)": 61.09, + "step": 6980, + "train_speed(iter/s)": 0.766773 + }, + { + "epoch": 0.6517690875232774, + "grad_norm": 1.7265625, + "learning_rate": 0.00018508965810738447, + "loss": 0.35780354, + "memory(GiB)": 61.09, + "step": 7000, + "train_speed(iter/s)": 0.766857 + }, + { + "epoch": 0.6536312849162011, + "grad_norm": 1.8515625, + "learning_rate": 0.00018498164245741555, + "loss": 0.34599066, + "memory(GiB)": 61.09, + "step": 7020, + "train_speed(iter/s)": 0.76696 + }, + { + "epoch": 0.6554934823091247, + "grad_norm": 2.171875, + "learning_rate": 0.00018487326872220505, + "loss": 0.37994871, + "memory(GiB)": 61.09, + "step": 7040, + "train_speed(iter/s)": 0.767115 + }, + { + "epoch": 0.6573556797020484, + "grad_norm": 1.75, + "learning_rate": 0.00018476453735840492, + "loss": 0.37087135, + "memory(GiB)": 61.09, + "step": 7060, + "train_speed(iter/s)": 0.76724 + }, + { + "epoch": 0.659217877094972, + "grad_norm": 2.375, + "learning_rate": 0.00018465544882417413, + "loss": 0.3895617, + "memory(GiB)": 61.09, + "step": 7080, + "train_speed(iter/s)": 0.767307 + }, + { + "epoch": 0.6610800744878957, + "grad_norm": 1.75, + "learning_rate": 0.00018454600357917656, + "loss": 0.36390953, + "memory(GiB)": 61.09, + "step": 7100, + "train_speed(iter/s)": 0.76745 + }, + { + "epoch": 0.6629422718808193, + "grad_norm": 1.7890625, + "learning_rate": 0.0001844362020845792, + "loss": 0.36717396, + "memory(GiB)": 61.09, + "step": 7120, + "train_speed(iter/s)": 0.767583 + }, + { + "epoch": 0.664804469273743, + "grad_norm": 2.0, + "learning_rate": 0.00018432604480305011, + "loss": 0.37814822, + "memory(GiB)": 61.09, + "step": 7140, + "train_speed(iter/s)": 0.767624 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 1.9375, + "learning_rate": 0.00018421553219875658, + "loss": 0.35424995, + "memory(GiB)": 61.09, + "step": 7160, + "train_speed(iter/s)": 0.767752 + }, + { + "epoch": 0.6685288640595903, + "grad_norm": 1.625, + "learning_rate": 0.00018410466473736312, + "loss": 0.36877232, + "memory(GiB)": 61.09, + "step": 7180, + "train_speed(iter/s)": 0.767815 + }, + { + "epoch": 0.6703910614525139, + "grad_norm": 1.6640625, + "learning_rate": 0.0001839934428860294, + "loss": 0.34407656, + "memory(GiB)": 61.09, + "step": 7200, + "train_speed(iter/s)": 0.767919 + }, + { + "epoch": 0.6722532588454376, + "grad_norm": 1.6640625, + "learning_rate": 0.00018388186711340852, + "loss": 0.38237391, + "memory(GiB)": 61.09, + "step": 7220, + "train_speed(iter/s)": 0.768033 + }, + { + "epoch": 0.6741154562383612, + "grad_norm": 2.046875, + "learning_rate": 0.0001837699378896448, + "loss": 0.3648797, + "memory(GiB)": 61.09, + "step": 7240, + "train_speed(iter/s)": 0.768127 + }, + { + "epoch": 0.6759776536312849, + "grad_norm": 1.6875, + "learning_rate": 0.00018365765568637178, + "loss": 0.33588786, + "memory(GiB)": 61.09, + "step": 7260, + "train_speed(iter/s)": 0.768202 + }, + { + "epoch": 0.6778398510242085, + "grad_norm": 1.515625, + "learning_rate": 0.00018354502097671062, + "loss": 0.3664793, + "memory(GiB)": 61.09, + "step": 7280, + "train_speed(iter/s)": 0.768262 + }, + { + "epoch": 0.6797020484171322, + "grad_norm": 1.75, + "learning_rate": 0.00018343203423526762, + "loss": 0.34410334, + "memory(GiB)": 61.09, + "step": 7300, + "train_speed(iter/s)": 0.768354 + }, + { + "epoch": 0.6815642458100558, + "grad_norm": 1.6953125, + "learning_rate": 0.00018331869593813248, + "loss": 0.35052714, + "memory(GiB)": 61.09, + "step": 7320, + "train_speed(iter/s)": 0.768408 + }, + { + "epoch": 0.6834264432029795, + "grad_norm": 1.8671875, + "learning_rate": 0.00018320500656287617, + "loss": 0.3694984, + "memory(GiB)": 61.09, + "step": 7340, + "train_speed(iter/s)": 0.768539 + }, + { + "epoch": 0.6852886405959032, + "grad_norm": 1.4375, + "learning_rate": 0.00018309096658854917, + "loss": 0.3540817, + "memory(GiB)": 61.09, + "step": 7360, + "train_speed(iter/s)": 0.76859 + }, + { + "epoch": 0.6871508379888268, + "grad_norm": 1.5546875, + "learning_rate": 0.0001829765764956791, + "loss": 0.32105043, + "memory(GiB)": 61.09, + "step": 7380, + "train_speed(iter/s)": 0.768719 + }, + { + "epoch": 0.6890130353817505, + "grad_norm": 1.21875, + "learning_rate": 0.00018286183676626894, + "loss": 0.36089745, + "memory(GiB)": 61.09, + "step": 7400, + "train_speed(iter/s)": 0.768705 + }, + { + "epoch": 0.6908752327746741, + "grad_norm": 1.5546875, + "learning_rate": 0.0001827467478837949, + "loss": 0.33547273, + "memory(GiB)": 61.09, + "step": 7420, + "train_speed(iter/s)": 0.768804 + }, + { + "epoch": 0.6927374301675978, + "grad_norm": 1.8046875, + "learning_rate": 0.00018263131033320447, + "loss": 0.32867005, + "memory(GiB)": 61.09, + "step": 7440, + "train_speed(iter/s)": 0.768897 + }, + { + "epoch": 0.6945996275605214, + "grad_norm": 1.859375, + "learning_rate": 0.00018251552460091423, + "loss": 0.34813719, + "memory(GiB)": 61.09, + "step": 7460, + "train_speed(iter/s)": 0.769038 + }, + { + "epoch": 0.6964618249534451, + "grad_norm": 1.8984375, + "learning_rate": 0.00018239939117480796, + "loss": 0.34680479, + "memory(GiB)": 61.09, + "step": 7480, + "train_speed(iter/s)": 0.769117 + }, + { + "epoch": 0.6983240223463687, + "grad_norm": 1.6796875, + "learning_rate": 0.00018228291054423444, + "loss": 0.34514585, + "memory(GiB)": 61.09, + "step": 7500, + "train_speed(iter/s)": 0.769217 + }, + { + "epoch": 0.7001862197392924, + "grad_norm": 1.703125, + "learning_rate": 0.00018216608320000547, + "loss": 0.36143909, + "memory(GiB)": 61.09, + "step": 7520, + "train_speed(iter/s)": 0.769262 + }, + { + "epoch": 0.702048417132216, + "grad_norm": 1.2734375, + "learning_rate": 0.00018204890963439388, + "loss": 0.34863019, + "memory(GiB)": 61.09, + "step": 7540, + "train_speed(iter/s)": 0.769313 + }, + { + "epoch": 0.7039106145251397, + "grad_norm": 1.4609375, + "learning_rate": 0.00018193139034113122, + "loss": 0.33534517, + "memory(GiB)": 61.09, + "step": 7560, + "train_speed(iter/s)": 0.769396 + }, + { + "epoch": 0.7057728119180633, + "grad_norm": 3.421875, + "learning_rate": 0.000181813525815406, + "loss": 0.37421381, + "memory(GiB)": 61.09, + "step": 7580, + "train_speed(iter/s)": 0.769524 + }, + { + "epoch": 0.707635009310987, + "grad_norm": 1.75, + "learning_rate": 0.0001816953165538612, + "loss": 0.35957568, + "memory(GiB)": 61.09, + "step": 7600, + "train_speed(iter/s)": 0.769611 + }, + { + "epoch": 0.7094972067039106, + "grad_norm": 1.8046875, + "learning_rate": 0.00018157676305459268, + "loss": 0.34478993, + "memory(GiB)": 61.09, + "step": 7620, + "train_speed(iter/s)": 0.769707 + }, + { + "epoch": 0.7113594040968343, + "grad_norm": 1.2890625, + "learning_rate": 0.00018145786581714654, + "loss": 0.34837961, + "memory(GiB)": 61.09, + "step": 7640, + "train_speed(iter/s)": 0.769876 + }, + { + "epoch": 0.7132216014897579, + "grad_norm": 1.3828125, + "learning_rate": 0.00018133862534251748, + "loss": 0.3299273, + "memory(GiB)": 61.09, + "step": 7660, + "train_speed(iter/s)": 0.770039 + }, + { + "epoch": 0.7150837988826816, + "grad_norm": 1.40625, + "learning_rate": 0.00018121904213314632, + "loss": 0.3494205, + "memory(GiB)": 61.09, + "step": 7680, + "train_speed(iter/s)": 0.770117 + }, + { + "epoch": 0.7169459962756052, + "grad_norm": 2.640625, + "learning_rate": 0.0001810991166929183, + "loss": 0.33510184, + "memory(GiB)": 61.09, + "step": 7700, + "train_speed(iter/s)": 0.77021 + }, + { + "epoch": 0.7188081936685289, + "grad_norm": 2.0625, + "learning_rate": 0.0001809788495271604, + "loss": 0.34384561, + "memory(GiB)": 61.09, + "step": 7720, + "train_speed(iter/s)": 0.770295 + }, + { + "epoch": 0.7206703910614525, + "grad_norm": 1.875, + "learning_rate": 0.00018085824114263981, + "loss": 0.35482507, + "memory(GiB)": 61.09, + "step": 7740, + "train_speed(iter/s)": 0.770398 + }, + { + "epoch": 0.7225325884543762, + "grad_norm": 1.296875, + "learning_rate": 0.0001807372920475613, + "loss": 0.3394609, + "memory(GiB)": 61.09, + "step": 7760, + "train_speed(iter/s)": 0.770488 + }, + { + "epoch": 0.7243947858472998, + "grad_norm": 1.453125, + "learning_rate": 0.0001806160027515654, + "loss": 0.3570467, + "memory(GiB)": 61.09, + "step": 7780, + "train_speed(iter/s)": 0.770601 + }, + { + "epoch": 0.7262569832402235, + "grad_norm": 1.5390625, + "learning_rate": 0.00018049437376572603, + "loss": 0.32628872, + "memory(GiB)": 61.09, + "step": 7800, + "train_speed(iter/s)": 0.77061 + }, + { + "epoch": 0.7281191806331471, + "grad_norm": 1.890625, + "learning_rate": 0.0001803724056025486, + "loss": 0.34269819, + "memory(GiB)": 61.09, + "step": 7820, + "train_speed(iter/s)": 0.770763 + }, + { + "epoch": 0.7299813780260708, + "grad_norm": 1.546875, + "learning_rate": 0.00018025009877596757, + "loss": 0.35294495, + "memory(GiB)": 61.09, + "step": 7840, + "train_speed(iter/s)": 0.770808 + }, + { + "epoch": 0.7318435754189944, + "grad_norm": 1.796875, + "learning_rate": 0.00018012745380134452, + "loss": 0.34454529, + "memory(GiB)": 61.09, + "step": 7860, + "train_speed(iter/s)": 0.770915 + }, + { + "epoch": 0.7337057728119181, + "grad_norm": 1.6171875, + "learning_rate": 0.0001800044711954658, + "loss": 0.32856019, + "memory(GiB)": 61.09, + "step": 7880, + "train_speed(iter/s)": 0.771015 + }, + { + "epoch": 0.7355679702048417, + "grad_norm": 1.2578125, + "learning_rate": 0.00017988115147654044, + "loss": 0.35125015, + "memory(GiB)": 61.09, + "step": 7900, + "train_speed(iter/s)": 0.771089 + }, + { + "epoch": 0.7374301675977654, + "grad_norm": 1.7890625, + "learning_rate": 0.000179757495164198, + "loss": 0.36670623, + "memory(GiB)": 61.09, + "step": 7920, + "train_speed(iter/s)": 0.771186 + }, + { + "epoch": 0.7392923649906891, + "grad_norm": 1.703125, + "learning_rate": 0.00017963350277948634, + "loss": 0.3566488, + "memory(GiB)": 61.09, + "step": 7940, + "train_speed(iter/s)": 0.771297 + }, + { + "epoch": 0.7411545623836127, + "grad_norm": 1.734375, + "learning_rate": 0.00017950917484486936, + "loss": 0.33071799, + "memory(GiB)": 61.09, + "step": 7960, + "train_speed(iter/s)": 0.771381 + }, + { + "epoch": 0.7430167597765364, + "grad_norm": 1.4609375, + "learning_rate": 0.00017938451188422488, + "loss": 0.34926617, + "memory(GiB)": 61.09, + "step": 7980, + "train_speed(iter/s)": 0.771511 + }, + { + "epoch": 0.74487895716946, + "grad_norm": 1.8515625, + "learning_rate": 0.00017925951442284243, + "loss": 0.36137218, + "memory(GiB)": 61.09, + "step": 8000, + "train_speed(iter/s)": 0.771625 + }, + { + "epoch": 0.74487895716946, + "eval_loss": 0.44029951095581055, + "eval_runtime": 76.4367, + "eval_samples_per_second": 181.654, + "eval_steps_per_second": 1.426, + "step": 8000 + }, + { + "epoch": 0.7467411545623837, + "grad_norm": 1.4609375, + "learning_rate": 0.00017913418298742104, + "loss": 0.34908738, + "memory(GiB)": 61.09, + "step": 8020, + "train_speed(iter/s)": 0.763602 + }, + { + "epoch": 0.7486033519553073, + "grad_norm": 1.4453125, + "learning_rate": 0.00017900851810606692, + "loss": 0.36671166, + "memory(GiB)": 61.09, + "step": 8040, + "train_speed(iter/s)": 0.763666 + }, + { + "epoch": 0.750465549348231, + "grad_norm": 1.8515625, + "learning_rate": 0.00017888252030829138, + "loss": 0.34246197, + "memory(GiB)": 61.09, + "step": 8060, + "train_speed(iter/s)": 0.763721 + }, + { + "epoch": 0.7523277467411545, + "grad_norm": 1.4921875, + "learning_rate": 0.00017875619012500856, + "loss": 0.33407593, + "memory(GiB)": 61.09, + "step": 8080, + "train_speed(iter/s)": 0.763808 + }, + { + "epoch": 0.7541899441340782, + "grad_norm": 1.734375, + "learning_rate": 0.00017862952808853305, + "loss": 0.32741477, + "memory(GiB)": 61.09, + "step": 8100, + "train_speed(iter/s)": 0.763914 + }, + { + "epoch": 0.7560521415270018, + "grad_norm": 1.3828125, + "learning_rate": 0.00017850253473257788, + "loss": 0.35220075, + "memory(GiB)": 61.09, + "step": 8120, + "train_speed(iter/s)": 0.76395 + }, + { + "epoch": 0.7579143389199255, + "grad_norm": 1.6875, + "learning_rate": 0.00017837521059225212, + "loss": 0.33165274, + "memory(GiB)": 61.09, + "step": 8140, + "train_speed(iter/s)": 0.7641 + }, + { + "epoch": 0.7597765363128491, + "grad_norm": 1.5859375, + "learning_rate": 0.0001782475562040586, + "loss": 0.36606779, + "memory(GiB)": 61.09, + "step": 8160, + "train_speed(iter/s)": 0.764186 + }, + { + "epoch": 0.7616387337057728, + "grad_norm": 1.7109375, + "learning_rate": 0.00017811957210589176, + "loss": 0.33115237, + "memory(GiB)": 61.09, + "step": 8180, + "train_speed(iter/s)": 0.764259 + }, + { + "epoch": 0.7635009310986964, + "grad_norm": 1.609375, + "learning_rate": 0.00017799125883703537, + "loss": 0.34380808, + "memory(GiB)": 61.09, + "step": 8200, + "train_speed(iter/s)": 0.764342 + }, + { + "epoch": 0.7653631284916201, + "grad_norm": 1.5078125, + "learning_rate": 0.00017786261693816013, + "loss": 0.34643345, + "memory(GiB)": 61.09, + "step": 8220, + "train_speed(iter/s)": 0.764458 + }, + { + "epoch": 0.7672253258845437, + "grad_norm": 1.5546875, + "learning_rate": 0.00017773364695132152, + "loss": 0.3174742, + "memory(GiB)": 61.09, + "step": 8240, + "train_speed(iter/s)": 0.764536 + }, + { + "epoch": 0.7690875232774674, + "grad_norm": 1.9296875, + "learning_rate": 0.00017760434941995753, + "loss": 0.35655365, + "memory(GiB)": 61.09, + "step": 8260, + "train_speed(iter/s)": 0.764581 + }, + { + "epoch": 0.770949720670391, + "grad_norm": 1.7578125, + "learning_rate": 0.00017747472488888622, + "loss": 0.3317188, + "memory(GiB)": 61.09, + "step": 8280, + "train_speed(iter/s)": 0.764703 + }, + { + "epoch": 0.7728119180633147, + "grad_norm": 1.4765625, + "learning_rate": 0.00017734477390430359, + "loss": 0.32900805, + "memory(GiB)": 61.09, + "step": 8300, + "train_speed(iter/s)": 0.764809 + }, + { + "epoch": 0.7746741154562383, + "grad_norm": 1.375, + "learning_rate": 0.0001772144970137812, + "loss": 0.33518133, + "memory(GiB)": 61.09, + "step": 8320, + "train_speed(iter/s)": 0.764899 + }, + { + "epoch": 0.776536312849162, + "grad_norm": 2.03125, + "learning_rate": 0.00017708389476626385, + "loss": 0.3434902, + "memory(GiB)": 61.09, + "step": 8340, + "train_speed(iter/s)": 0.764975 + }, + { + "epoch": 0.7783985102420856, + "grad_norm": 1.6640625, + "learning_rate": 0.00017695296771206732, + "loss": 0.32565141, + "memory(GiB)": 61.09, + "step": 8360, + "train_speed(iter/s)": 0.765046 + }, + { + "epoch": 0.7802607076350093, + "grad_norm": 1.3046875, + "learning_rate": 0.00017682171640287598, + "loss": 0.33638844, + "memory(GiB)": 61.09, + "step": 8380, + "train_speed(iter/s)": 0.765188 + }, + { + "epoch": 0.7821229050279329, + "grad_norm": 1.0546875, + "learning_rate": 0.00017669014139174054, + "loss": 0.33523929, + "memory(GiB)": 61.09, + "step": 8400, + "train_speed(iter/s)": 0.765285 + }, + { + "epoch": 0.7839851024208566, + "grad_norm": 1.625, + "learning_rate": 0.0001765582432330757, + "loss": 0.34908547, + "memory(GiB)": 61.09, + "step": 8420, + "train_speed(iter/s)": 0.765416 + }, + { + "epoch": 0.7858472998137802, + "grad_norm": 1.53125, + "learning_rate": 0.00017642602248265767, + "loss": 0.31645274, + "memory(GiB)": 61.09, + "step": 8440, + "train_speed(iter/s)": 0.765534 + }, + { + "epoch": 0.7877094972067039, + "grad_norm": 1.484375, + "learning_rate": 0.0001762934796976222, + "loss": 0.34835916, + "memory(GiB)": 61.09, + "step": 8460, + "train_speed(iter/s)": 0.765627 + }, + { + "epoch": 0.7895716945996276, + "grad_norm": 1.5390625, + "learning_rate": 0.00017616061543646177, + "loss": 0.34466844, + "memory(GiB)": 61.09, + "step": 8480, + "train_speed(iter/s)": 0.765714 + }, + { + "epoch": 0.7914338919925512, + "grad_norm": 1.6796875, + "learning_rate": 0.00017602743025902355, + "loss": 0.34232378, + "memory(GiB)": 61.09, + "step": 8500, + "train_speed(iter/s)": 0.765833 + }, + { + "epoch": 0.7932960893854749, + "grad_norm": 1.453125, + "learning_rate": 0.00017589392472650697, + "loss": 0.33207898, + "memory(GiB)": 61.09, + "step": 8520, + "train_speed(iter/s)": 0.765895 + }, + { + "epoch": 0.7951582867783985, + "grad_norm": 1.2265625, + "learning_rate": 0.00017576009940146128, + "loss": 0.31665926, + "memory(GiB)": 61.09, + "step": 8540, + "train_speed(iter/s)": 0.765888 + }, + { + "epoch": 0.7970204841713222, + "grad_norm": 1.7265625, + "learning_rate": 0.00017562595484778322, + "loss": 0.33440113, + "memory(GiB)": 61.09, + "step": 8560, + "train_speed(iter/s)": 0.766027 + }, + { + "epoch": 0.7988826815642458, + "grad_norm": 2.328125, + "learning_rate": 0.0001754914916307147, + "loss": 0.34350667, + "memory(GiB)": 61.09, + "step": 8580, + "train_speed(iter/s)": 0.766145 + }, + { + "epoch": 0.8007448789571695, + "grad_norm": 1.359375, + "learning_rate": 0.00017535671031684041, + "loss": 0.33459611, + "memory(GiB)": 61.09, + "step": 8600, + "train_speed(iter/s)": 0.76629 + }, + { + "epoch": 0.8026070763500931, + "grad_norm": 1.6015625, + "learning_rate": 0.00017522161147408532, + "loss": 0.34758761, + "memory(GiB)": 61.09, + "step": 8620, + "train_speed(iter/s)": 0.766406 + }, + { + "epoch": 0.8044692737430168, + "grad_norm": 1.640625, + "learning_rate": 0.0001750861956717124, + "loss": 0.31787, + "memory(GiB)": 61.09, + "step": 8640, + "train_speed(iter/s)": 0.766498 + }, + { + "epoch": 0.8063314711359404, + "grad_norm": 1.65625, + "learning_rate": 0.00017495046348032015, + "loss": 0.3213912, + "memory(GiB)": 61.09, + "step": 8660, + "train_speed(iter/s)": 0.766583 + }, + { + "epoch": 0.8081936685288641, + "grad_norm": 1.6796875, + "learning_rate": 0.0001748144154718403, + "loss": 0.34279273, + "memory(GiB)": 61.09, + "step": 8680, + "train_speed(iter/s)": 0.766674 + }, + { + "epoch": 0.8100558659217877, + "grad_norm": 1.59375, + "learning_rate": 0.00017467805221953525, + "loss": 0.30648351, + "memory(GiB)": 61.09, + "step": 8700, + "train_speed(iter/s)": 0.76673 + }, + { + "epoch": 0.8119180633147114, + "grad_norm": 1.6484375, + "learning_rate": 0.00017454137429799582, + "loss": 0.34818337, + "memory(GiB)": 61.09, + "step": 8720, + "train_speed(iter/s)": 0.766821 + }, + { + "epoch": 0.813780260707635, + "grad_norm": 1.8515625, + "learning_rate": 0.0001744043822831387, + "loss": 0.35243084, + "memory(GiB)": 61.09, + "step": 8740, + "train_speed(iter/s)": 0.766816 + }, + { + "epoch": 0.8156424581005587, + "grad_norm": 1.859375, + "learning_rate": 0.00017426707675220407, + "loss": 0.3365181, + "memory(GiB)": 61.09, + "step": 8760, + "train_speed(iter/s)": 0.766823 + }, + { + "epoch": 0.8175046554934823, + "grad_norm": 2.0625, + "learning_rate": 0.00017412945828375315, + "loss": 0.30240817, + "memory(GiB)": 61.09, + "step": 8780, + "train_speed(iter/s)": 0.766921 + }, + { + "epoch": 0.819366852886406, + "grad_norm": 1.3828125, + "learning_rate": 0.00017399152745766584, + "loss": 0.32030184, + "memory(GiB)": 61.09, + "step": 8800, + "train_speed(iter/s)": 0.767073 + }, + { + "epoch": 0.8212290502793296, + "grad_norm": 1.765625, + "learning_rate": 0.00017385328485513806, + "loss": 0.3264648, + "memory(GiB)": 61.09, + "step": 8820, + "train_speed(iter/s)": 0.767141 + }, + { + "epoch": 0.8230912476722533, + "grad_norm": 1.5234375, + "learning_rate": 0.0001737147310586797, + "loss": 0.32821541, + "memory(GiB)": 61.09, + "step": 8840, + "train_speed(iter/s)": 0.76719 + }, + { + "epoch": 0.8249534450651769, + "grad_norm": 1.4765625, + "learning_rate": 0.0001735758666521117, + "loss": 0.35671031, + "memory(GiB)": 61.09, + "step": 8860, + "train_speed(iter/s)": 0.767275 + }, + { + "epoch": 0.8268156424581006, + "grad_norm": 1.5234375, + "learning_rate": 0.00017343669222056395, + "loss": 0.33457453, + "memory(GiB)": 61.09, + "step": 8880, + "train_speed(iter/s)": 0.767361 + }, + { + "epoch": 0.8286778398510242, + "grad_norm": 1.6640625, + "learning_rate": 0.00017329720835047255, + "loss": 0.32296019, + "memory(GiB)": 61.09, + "step": 8900, + "train_speed(iter/s)": 0.767461 + }, + { + "epoch": 0.8305400372439479, + "grad_norm": 1.2890625, + "learning_rate": 0.00017315741562957767, + "loss": 0.32958663, + "memory(GiB)": 61.09, + "step": 8920, + "train_speed(iter/s)": 0.767551 + }, + { + "epoch": 0.8324022346368715, + "grad_norm": 1.234375, + "learning_rate": 0.0001730173146469207, + "loss": 0.32108679, + "memory(GiB)": 61.09, + "step": 8940, + "train_speed(iter/s)": 0.767576 + }, + { + "epoch": 0.8342644320297952, + "grad_norm": 1.578125, + "learning_rate": 0.00017287690599284198, + "loss": 0.32883315, + "memory(GiB)": 61.09, + "step": 8960, + "train_speed(iter/s)": 0.767694 + }, + { + "epoch": 0.8361266294227188, + "grad_norm": 1.421875, + "learning_rate": 0.00017273619025897835, + "loss": 0.33148849, + "memory(GiB)": 61.09, + "step": 8980, + "train_speed(iter/s)": 0.767813 + }, + { + "epoch": 0.8379888268156425, + "grad_norm": 2.09375, + "learning_rate": 0.00017259516803826052, + "loss": 0.32300215, + "memory(GiB)": 61.09, + "step": 9000, + "train_speed(iter/s)": 0.767839 + }, + { + "epoch": 0.839851024208566, + "grad_norm": 1.578125, + "learning_rate": 0.00017245383992491072, + "loss": 0.3056927, + "memory(GiB)": 61.09, + "step": 9020, + "train_speed(iter/s)": 0.767908 + }, + { + "epoch": 0.8417132216014898, + "grad_norm": 1.8203125, + "learning_rate": 0.00017231220651443997, + "loss": 0.32388425, + "memory(GiB)": 61.09, + "step": 9040, + "train_speed(iter/s)": 0.767979 + }, + { + "epoch": 0.8435754189944135, + "grad_norm": 1.234375, + "learning_rate": 0.00017217026840364588, + "loss": 0.32691157, + "memory(GiB)": 61.09, + "step": 9060, + "train_speed(iter/s)": 0.768042 + }, + { + "epoch": 0.845437616387337, + "grad_norm": 1.4296875, + "learning_rate": 0.0001720280261906098, + "loss": 0.31360347, + "memory(GiB)": 61.09, + "step": 9080, + "train_speed(iter/s)": 0.768054 + }, + { + "epoch": 0.8472998137802608, + "grad_norm": 1.4921875, + "learning_rate": 0.0001718854804746946, + "loss": 0.33109579, + "memory(GiB)": 61.09, + "step": 9100, + "train_speed(iter/s)": 0.76809 + }, + { + "epoch": 0.8491620111731844, + "grad_norm": 1.6796875, + "learning_rate": 0.00017174263185654195, + "loss": 0.315205, + "memory(GiB)": 61.09, + "step": 9120, + "train_speed(iter/s)": 0.768076 + }, + { + "epoch": 0.851024208566108, + "grad_norm": 2.15625, + "learning_rate": 0.00017159948093806987, + "loss": 0.33985429, + "memory(GiB)": 61.09, + "step": 9140, + "train_speed(iter/s)": 0.768161 + }, + { + "epoch": 0.8528864059590316, + "grad_norm": 1.4375, + "learning_rate": 0.00017145602832247024, + "loss": 0.32494226, + "memory(GiB)": 61.09, + "step": 9160, + "train_speed(iter/s)": 0.768242 + }, + { + "epoch": 0.8547486033519553, + "grad_norm": 1.46875, + "learning_rate": 0.00017131227461420604, + "loss": 0.33872006, + "memory(GiB)": 61.09, + "step": 9180, + "train_speed(iter/s)": 0.76826 + }, + { + "epoch": 0.8566108007448789, + "grad_norm": 1.4765625, + "learning_rate": 0.00017116822041900908, + "loss": 0.33576746, + "memory(GiB)": 61.09, + "step": 9200, + "train_speed(iter/s)": 0.768325 + }, + { + "epoch": 0.8584729981378026, + "grad_norm": 1.53125, + "learning_rate": 0.00017102386634387733, + "loss": 0.32657223, + "memory(GiB)": 61.09, + "step": 9220, + "train_speed(iter/s)": 0.768388 + }, + { + "epoch": 0.8603351955307262, + "grad_norm": 1.4765625, + "learning_rate": 0.00017087921299707227, + "loss": 0.32885175, + "memory(GiB)": 61.09, + "step": 9240, + "train_speed(iter/s)": 0.768406 + }, + { + "epoch": 0.8621973929236499, + "grad_norm": 1.6484375, + "learning_rate": 0.00017073426098811645, + "loss": 0.3327816, + "memory(GiB)": 61.09, + "step": 9260, + "train_speed(iter/s)": 0.768456 + }, + { + "epoch": 0.8640595903165735, + "grad_norm": 1.6171875, + "learning_rate": 0.00017058901092779097, + "loss": 0.31467597, + "memory(GiB)": 61.09, + "step": 9280, + "train_speed(iter/s)": 0.768571 + }, + { + "epoch": 0.8659217877094972, + "grad_norm": 1.3984375, + "learning_rate": 0.0001704434634281327, + "loss": 0.33298354, + "memory(GiB)": 61.09, + "step": 9300, + "train_speed(iter/s)": 0.76864 + }, + { + "epoch": 0.8677839851024208, + "grad_norm": 1.375, + "learning_rate": 0.00017029761910243186, + "loss": 0.33006821, + "memory(GiB)": 61.09, + "step": 9320, + "train_speed(iter/s)": 0.768736 + }, + { + "epoch": 0.8696461824953445, + "grad_norm": 1.5078125, + "learning_rate": 0.0001701514785652295, + "loss": 0.32524149, + "memory(GiB)": 61.09, + "step": 9340, + "train_speed(iter/s)": 0.768775 + }, + { + "epoch": 0.8715083798882681, + "grad_norm": 1.3515625, + "learning_rate": 0.00017000504243231464, + "loss": 0.31099353, + "memory(GiB)": 61.09, + "step": 9360, + "train_speed(iter/s)": 0.768862 + }, + { + "epoch": 0.8733705772811918, + "grad_norm": 1.7578125, + "learning_rate": 0.00016985831132072197, + "loss": 0.35106015, + "memory(GiB)": 61.09, + "step": 9380, + "train_speed(iter/s)": 0.768879 + }, + { + "epoch": 0.8752327746741154, + "grad_norm": 1.7265625, + "learning_rate": 0.0001697112858487291, + "loss": 0.34133174, + "memory(GiB)": 61.09, + "step": 9400, + "train_speed(iter/s)": 0.768871 + }, + { + "epoch": 0.8770949720670391, + "grad_norm": 1.7265625, + "learning_rate": 0.0001695639666358539, + "loss": 0.32821767, + "memory(GiB)": 61.09, + "step": 9420, + "train_speed(iter/s)": 0.768959 + }, + { + "epoch": 0.8789571694599627, + "grad_norm": 1.375, + "learning_rate": 0.00016941635430285206, + "loss": 0.30042934, + "memory(GiB)": 61.09, + "step": 9440, + "train_speed(iter/s)": 0.769071 + }, + { + "epoch": 0.8808193668528864, + "grad_norm": 1.765625, + "learning_rate": 0.00016926844947171444, + "loss": 0.30185919, + "memory(GiB)": 61.09, + "step": 9460, + "train_speed(iter/s)": 0.769178 + }, + { + "epoch": 0.88268156424581, + "grad_norm": 1.484375, + "learning_rate": 0.0001691202527656642, + "loss": 0.32123606, + "memory(GiB)": 61.09, + "step": 9480, + "train_speed(iter/s)": 0.76926 + }, + { + "epoch": 0.8845437616387337, + "grad_norm": 1.90625, + "learning_rate": 0.00016897176480915453, + "loss": 0.33448987, + "memory(GiB)": 61.09, + "step": 9500, + "train_speed(iter/s)": 0.769379 + }, + { + "epoch": 0.8864059590316573, + "grad_norm": 1.578125, + "learning_rate": 0.00016882298622786582, + "loss": 0.32368143, + "memory(GiB)": 61.09, + "step": 9520, + "train_speed(iter/s)": 0.769475 + }, + { + "epoch": 0.888268156424581, + "grad_norm": 1.34375, + "learning_rate": 0.000168673917648703, + "loss": 0.28515136, + "memory(GiB)": 61.09, + "step": 9540, + "train_speed(iter/s)": 0.769458 + }, + { + "epoch": 0.8901303538175046, + "grad_norm": 1.75, + "learning_rate": 0.00016852455969979302, + "loss": 0.32660358, + "memory(GiB)": 61.09, + "step": 9560, + "train_speed(iter/s)": 0.769559 + }, + { + "epoch": 0.8919925512104283, + "grad_norm": 1.7578125, + "learning_rate": 0.00016837491301048213, + "loss": 0.31508813, + "memory(GiB)": 61.09, + "step": 9580, + "train_speed(iter/s)": 0.769661 + }, + { + "epoch": 0.8938547486033519, + "grad_norm": 1.671875, + "learning_rate": 0.00016822497821133322, + "loss": 0.3051513, + "memory(GiB)": 61.09, + "step": 9600, + "train_speed(iter/s)": 0.769764 + }, + { + "epoch": 0.8957169459962756, + "grad_norm": 1.2421875, + "learning_rate": 0.00016807475593412315, + "loss": 0.29644897, + "memory(GiB)": 61.09, + "step": 9620, + "train_speed(iter/s)": 0.769809 + }, + { + "epoch": 0.8975791433891993, + "grad_norm": 1.375, + "learning_rate": 0.00016792424681184024, + "loss": 0.327367, + "memory(GiB)": 61.09, + "step": 9640, + "train_speed(iter/s)": 0.76979 + }, + { + "epoch": 0.8994413407821229, + "grad_norm": 1.59375, + "learning_rate": 0.00016777345147868137, + "loss": 0.28873682, + "memory(GiB)": 61.09, + "step": 9660, + "train_speed(iter/s)": 0.769893 + }, + { + "epoch": 0.9013035381750466, + "grad_norm": 1.671875, + "learning_rate": 0.00016762237057004945, + "loss": 0.3337399, + "memory(GiB)": 61.09, + "step": 9680, + "train_speed(iter/s)": 0.769961 + }, + { + "epoch": 0.9031657355679702, + "grad_norm": 1.4140625, + "learning_rate": 0.0001674710047225507, + "loss": 0.31436739, + "memory(GiB)": 61.09, + "step": 9700, + "train_speed(iter/s)": 0.770068 + }, + { + "epoch": 0.9050279329608939, + "grad_norm": 1.5625, + "learning_rate": 0.00016731935457399205, + "loss": 0.32119958, + "memory(GiB)": 61.09, + "step": 9720, + "train_speed(iter/s)": 0.770142 + }, + { + "epoch": 0.9068901303538175, + "grad_norm": 1.53125, + "learning_rate": 0.00016716742076337826, + "loss": 0.31439784, + "memory(GiB)": 61.09, + "step": 9740, + "train_speed(iter/s)": 0.7702 + }, + { + "epoch": 0.9087523277467412, + "grad_norm": 1.5859375, + "learning_rate": 0.00016701520393090947, + "loss": 0.30380776, + "memory(GiB)": 61.09, + "step": 9760, + "train_speed(iter/s)": 0.770272 + }, + { + "epoch": 0.9106145251396648, + "grad_norm": 1.921875, + "learning_rate": 0.00016686270471797832, + "loss": 0.32591763, + "memory(GiB)": 61.09, + "step": 9780, + "train_speed(iter/s)": 0.770365 + }, + { + "epoch": 0.9124767225325885, + "grad_norm": 1.6328125, + "learning_rate": 0.00016670992376716735, + "loss": 0.31843548, + "memory(GiB)": 61.09, + "step": 9800, + "train_speed(iter/s)": 0.770333 + }, + { + "epoch": 0.9143389199255121, + "grad_norm": 1.6484375, + "learning_rate": 0.0001665568617222462, + "loss": 0.33044133, + "memory(GiB)": 61.09, + "step": 9820, + "train_speed(iter/s)": 0.770405 + }, + { + "epoch": 0.9162011173184358, + "grad_norm": 2.09375, + "learning_rate": 0.00016640351922816898, + "loss": 0.30755372, + "memory(GiB)": 61.09, + "step": 9840, + "train_speed(iter/s)": 0.770498 + }, + { + "epoch": 0.9180633147113594, + "grad_norm": 1.6640625, + "learning_rate": 0.0001662498969310715, + "loss": 0.31706464, + "memory(GiB)": 61.09, + "step": 9860, + "train_speed(iter/s)": 0.770551 + }, + { + "epoch": 0.9199255121042831, + "grad_norm": 1.65625, + "learning_rate": 0.00016609599547826864, + "loss": 0.32224698, + "memory(GiB)": 61.09, + "step": 9880, + "train_speed(iter/s)": 0.770596 + }, + { + "epoch": 0.9217877094972067, + "grad_norm": 1.9453125, + "learning_rate": 0.00016594181551825147, + "loss": 0.3317049, + "memory(GiB)": 61.09, + "step": 9900, + "train_speed(iter/s)": 0.770655 + }, + { + "epoch": 0.9236499068901304, + "grad_norm": 1.859375, + "learning_rate": 0.00016578735770068464, + "loss": 0.28957124, + "memory(GiB)": 61.09, + "step": 9920, + "train_speed(iter/s)": 0.7707 + }, + { + "epoch": 0.925512104283054, + "grad_norm": 2.203125, + "learning_rate": 0.00016563262267640355, + "loss": 0.31882176, + "memory(GiB)": 61.09, + "step": 9940, + "train_speed(iter/s)": 0.77078 + }, + { + "epoch": 0.9273743016759777, + "grad_norm": 1.234375, + "learning_rate": 0.00016547761109741177, + "loss": 0.29138837, + "memory(GiB)": 61.09, + "step": 9960, + "train_speed(iter/s)": 0.770913 + }, + { + "epoch": 0.9292364990689013, + "grad_norm": 1.5234375, + "learning_rate": 0.00016532232361687806, + "loss": 0.32336102, + "memory(GiB)": 61.09, + "step": 9980, + "train_speed(iter/s)": 0.771 + }, + { + "epoch": 0.931098696461825, + "grad_norm": 1.359375, + "learning_rate": 0.00016516676088913386, + "loss": 0.30521989, + "memory(GiB)": 61.09, + "step": 10000, + "train_speed(iter/s)": 0.771064 + }, + { + "epoch": 0.931098696461825, + "eval_loss": 0.4275904893875122, + "eval_runtime": 76.1303, + "eval_samples_per_second": 182.385, + "eval_steps_per_second": 1.432, + "step": 10000 + }, + { + "epoch": 0.9329608938547486, + "grad_norm": 1.4296875, + "learning_rate": 0.0001650109235696703, + "loss": 0.3077956, + "memory(GiB)": 61.09, + "step": 10020, + "train_speed(iter/s)": 0.764779 + }, + { + "epoch": 0.9348230912476723, + "grad_norm": 1.5234375, + "learning_rate": 0.00016485481231513557, + "loss": 0.30179648, + "memory(GiB)": 61.09, + "step": 10040, + "train_speed(iter/s)": 0.764916 + }, + { + "epoch": 0.9366852886405959, + "grad_norm": 1.3359375, + "learning_rate": 0.00016469842778333224, + "loss": 0.29367597, + "memory(GiB)": 61.09, + "step": 10060, + "train_speed(iter/s)": 0.764969 + }, + { + "epoch": 0.9385474860335196, + "grad_norm": 1.578125, + "learning_rate": 0.00016454177063321423, + "loss": 0.32096641, + "memory(GiB)": 61.09, + "step": 10080, + "train_speed(iter/s)": 0.765074 + }, + { + "epoch": 0.9404096834264432, + "grad_norm": 1.3828125, + "learning_rate": 0.00016438484152488437, + "loss": 0.33787951, + "memory(GiB)": 61.09, + "step": 10100, + "train_speed(iter/s)": 0.765195 + }, + { + "epoch": 0.9422718808193669, + "grad_norm": 1.6796875, + "learning_rate": 0.00016422764111959117, + "loss": 0.33140016, + "memory(GiB)": 61.09, + "step": 10120, + "train_speed(iter/s)": 0.765314 + }, + { + "epoch": 0.9441340782122905, + "grad_norm": 1.3359375, + "learning_rate": 0.00016407017007972655, + "loss": 0.3267303, + "memory(GiB)": 61.09, + "step": 10140, + "train_speed(iter/s)": 0.765392 + }, + { + "epoch": 0.9459962756052142, + "grad_norm": 1.609375, + "learning_rate": 0.0001639124290688227, + "loss": 0.30506415, + "memory(GiB)": 61.09, + "step": 10160, + "train_speed(iter/s)": 0.765479 + }, + { + "epoch": 0.9478584729981379, + "grad_norm": 1.4140625, + "learning_rate": 0.0001637544187515494, + "loss": 0.32266409, + "memory(GiB)": 61.09, + "step": 10180, + "train_speed(iter/s)": 0.765539 + }, + { + "epoch": 0.9497206703910615, + "grad_norm": 1.1796875, + "learning_rate": 0.00016359613979371112, + "loss": 0.31371219, + "memory(GiB)": 61.09, + "step": 10200, + "train_speed(iter/s)": 0.765598 + }, + { + "epoch": 0.9515828677839852, + "grad_norm": 1.53125, + "learning_rate": 0.00016343759286224438, + "loss": 0.32517624, + "memory(GiB)": 61.09, + "step": 10220, + "train_speed(iter/s)": 0.765643 + }, + { + "epoch": 0.9534450651769087, + "grad_norm": 1.671875, + "learning_rate": 0.00016327877862521483, + "loss": 0.29492621, + "memory(GiB)": 61.09, + "step": 10240, + "train_speed(iter/s)": 0.765654 + }, + { + "epoch": 0.9553072625698324, + "grad_norm": 1.453125, + "learning_rate": 0.00016311969775181448, + "loss": 0.31768193, + "memory(GiB)": 61.09, + "step": 10260, + "train_speed(iter/s)": 0.765711 + }, + { + "epoch": 0.957169459962756, + "grad_norm": 1.5546875, + "learning_rate": 0.00016296035091235882, + "loss": 0.31737647, + "memory(GiB)": 61.09, + "step": 10280, + "train_speed(iter/s)": 0.765782 + }, + { + "epoch": 0.9590316573556797, + "grad_norm": 1.6328125, + "learning_rate": 0.00016280073877828407, + "loss": 0.29982419, + "memory(GiB)": 61.09, + "step": 10300, + "train_speed(iter/s)": 0.765856 + }, + { + "epoch": 0.9608938547486033, + "grad_norm": 1.3828125, + "learning_rate": 0.0001626408620221443, + "loss": 0.29922385, + "memory(GiB)": 61.09, + "step": 10320, + "train_speed(iter/s)": 0.765951 + }, + { + "epoch": 0.962756052141527, + "grad_norm": 1.3359375, + "learning_rate": 0.0001624807213176086, + "loss": 0.31153214, + "memory(GiB)": 61.09, + "step": 10340, + "train_speed(iter/s)": 0.765997 + }, + { + "epoch": 0.9646182495344506, + "grad_norm": 1.3671875, + "learning_rate": 0.00016232031733945824, + "loss": 0.32357917, + "memory(GiB)": 61.09, + "step": 10360, + "train_speed(iter/s)": 0.766017 + }, + { + "epoch": 0.9664804469273743, + "grad_norm": 1.5, + "learning_rate": 0.0001621596507635839, + "loss": 0.3050452, + "memory(GiB)": 61.09, + "step": 10380, + "train_speed(iter/s)": 0.766116 + }, + { + "epoch": 0.9683426443202979, + "grad_norm": 1.3671875, + "learning_rate": 0.00016199872226698277, + "loss": 0.30685663, + "memory(GiB)": 61.09, + "step": 10400, + "train_speed(iter/s)": 0.766219 + }, + { + "epoch": 0.9702048417132216, + "grad_norm": 1.4609375, + "learning_rate": 0.00016183753252775558, + "loss": 0.31804857, + "memory(GiB)": 61.09, + "step": 10420, + "train_speed(iter/s)": 0.766245 + }, + { + "epoch": 0.9720670391061452, + "grad_norm": 1.3203125, + "learning_rate": 0.00016167608222510392, + "loss": 0.28163512, + "memory(GiB)": 61.09, + "step": 10440, + "train_speed(iter/s)": 0.766326 + }, + { + "epoch": 0.9739292364990689, + "grad_norm": 1.5546875, + "learning_rate": 0.00016151437203932738, + "loss": 0.31363811, + "memory(GiB)": 61.09, + "step": 10460, + "train_speed(iter/s)": 0.766343 + }, + { + "epoch": 0.9757914338919925, + "grad_norm": 1.6484375, + "learning_rate": 0.00016135240265182046, + "loss": 0.31705124, + "memory(GiB)": 61.09, + "step": 10480, + "train_speed(iter/s)": 0.766417 + }, + { + "epoch": 0.9776536312849162, + "grad_norm": 1.6796875, + "learning_rate": 0.00016119017474507002, + "loss": 0.3048178, + "memory(GiB)": 61.09, + "step": 10500, + "train_speed(iter/s)": 0.766326 + }, + { + "epoch": 0.9795158286778398, + "grad_norm": 1.1875, + "learning_rate": 0.00016102768900265208, + "loss": 0.2878325, + "memory(GiB)": 61.09, + "step": 10520, + "train_speed(iter/s)": 0.766347 + }, + { + "epoch": 0.9813780260707635, + "grad_norm": 1.5, + "learning_rate": 0.00016086494610922917, + "loss": 0.30582104, + "memory(GiB)": 61.09, + "step": 10540, + "train_speed(iter/s)": 0.766427 + }, + { + "epoch": 0.9832402234636871, + "grad_norm": 1.703125, + "learning_rate": 0.00016070194675054747, + "loss": 0.30753689, + "memory(GiB)": 61.09, + "step": 10560, + "train_speed(iter/s)": 0.766485 + }, + { + "epoch": 0.9851024208566108, + "grad_norm": 1.4609375, + "learning_rate": 0.00016053869161343362, + "loss": 0.30274782, + "memory(GiB)": 61.09, + "step": 10580, + "train_speed(iter/s)": 0.766535 + }, + { + "epoch": 0.9869646182495344, + "grad_norm": 1.765625, + "learning_rate": 0.00016037518138579214, + "loss": 0.32131209, + "memory(GiB)": 61.09, + "step": 10600, + "train_speed(iter/s)": 0.766567 + }, + { + "epoch": 0.9888268156424581, + "grad_norm": 1.5703125, + "learning_rate": 0.00016021141675660248, + "loss": 0.30618265, + "memory(GiB)": 61.09, + "step": 10620, + "train_speed(iter/s)": 0.766659 + }, + { + "epoch": 0.9906890130353817, + "grad_norm": 1.859375, + "learning_rate": 0.00016004739841591592, + "loss": 0.29824829, + "memory(GiB)": 61.09, + "step": 10640, + "train_speed(iter/s)": 0.76669 + }, + { + "epoch": 0.9925512104283054, + "grad_norm": 1.75, + "learning_rate": 0.0001598831270548529, + "loss": 0.3051553, + "memory(GiB)": 61.09, + "step": 10660, + "train_speed(iter/s)": 0.766785 + }, + { + "epoch": 0.994413407821229, + "grad_norm": 1.5703125, + "learning_rate": 0.00015971860336559996, + "loss": 0.30366507, + "memory(GiB)": 61.09, + "step": 10680, + "train_speed(iter/s)": 0.766815 + }, + { + "epoch": 0.9962756052141527, + "grad_norm": 1.515625, + "learning_rate": 0.00015955382804140687, + "loss": 0.31902528, + "memory(GiB)": 61.09, + "step": 10700, + "train_speed(iter/s)": 0.76688 + }, + { + "epoch": 0.9981378026070763, + "grad_norm": 1.578125, + "learning_rate": 0.00015938880177658374, + "loss": 0.30881336, + "memory(GiB)": 61.09, + "step": 10720, + "train_speed(iter/s)": 0.766982 + }, + { + "epoch": 1.0, + "grad_norm": 1.453125, + "learning_rate": 0.00015922352526649803, + "loss": 0.30783999, + "memory(GiB)": 61.09, + "step": 10740, + "train_speed(iter/s)": 0.767075 + }, + { + "epoch": 1.0018621973929236, + "grad_norm": 1.296875, + "learning_rate": 0.00015905799920757163, + "loss": 0.2917798, + "memory(GiB)": 61.09, + "step": 10760, + "train_speed(iter/s)": 0.766514 + }, + { + "epoch": 1.0037243947858474, + "grad_norm": 1.7109375, + "learning_rate": 0.00015889222429727807, + "loss": 0.2846385, + "memory(GiB)": 61.09, + "step": 10780, + "train_speed(iter/s)": 0.766588 + }, + { + "epoch": 1.005586592178771, + "grad_norm": 1.4765625, + "learning_rate": 0.00015872620123413928, + "loss": 0.29505885, + "memory(GiB)": 61.09, + "step": 10800, + "train_speed(iter/s)": 0.76667 + }, + { + "epoch": 1.0074487895716946, + "grad_norm": 1.53125, + "learning_rate": 0.000158559930717723, + "loss": 0.27733662, + "memory(GiB)": 61.09, + "step": 10820, + "train_speed(iter/s)": 0.76677 + }, + { + "epoch": 1.0093109869646182, + "grad_norm": 1.6015625, + "learning_rate": 0.00015839341344863953, + "loss": 0.29862165, + "memory(GiB)": 61.09, + "step": 10840, + "train_speed(iter/s)": 0.766803 + }, + { + "epoch": 1.011173184357542, + "grad_norm": 1.7734375, + "learning_rate": 0.00015822665012853895, + "loss": 0.3004262, + "memory(GiB)": 61.09, + "step": 10860, + "train_speed(iter/s)": 0.766875 + }, + { + "epoch": 1.0130353817504656, + "grad_norm": 1.75, + "learning_rate": 0.00015805964146010815, + "loss": 0.32919612, + "memory(GiB)": 61.09, + "step": 10880, + "train_speed(iter/s)": 0.766929 + }, + { + "epoch": 1.0148975791433892, + "grad_norm": 1.4609375, + "learning_rate": 0.0001578923881470678, + "loss": 0.31359372, + "memory(GiB)": 61.09, + "step": 10900, + "train_speed(iter/s)": 0.766932 + }, + { + "epoch": 1.0167597765363128, + "grad_norm": 1.4765625, + "learning_rate": 0.00015772489089416943, + "loss": 0.3069638, + "memory(GiB)": 61.09, + "step": 10920, + "train_speed(iter/s)": 0.767026 + }, + { + "epoch": 1.0186219739292366, + "grad_norm": 1.1171875, + "learning_rate": 0.00015755715040719243, + "loss": 0.29340148, + "memory(GiB)": 61.09, + "step": 10940, + "train_speed(iter/s)": 0.7671 + }, + { + "epoch": 1.0204841713221602, + "grad_norm": 1.7109375, + "learning_rate": 0.00015738916739294122, + "loss": 0.29378917, + "memory(GiB)": 61.09, + "step": 10960, + "train_speed(iter/s)": 0.767181 + }, + { + "epoch": 1.0223463687150838, + "grad_norm": 1.5859375, + "learning_rate": 0.000157220942559242, + "loss": 0.30050142, + "memory(GiB)": 61.09, + "step": 10980, + "train_speed(iter/s)": 0.767282 + }, + { + "epoch": 1.0242085661080074, + "grad_norm": 1.6953125, + "learning_rate": 0.00015705247661493992, + "loss": 0.28569994, + "memory(GiB)": 61.09, + "step": 11000, + "train_speed(iter/s)": 0.767289 + }, + { + "epoch": 1.0260707635009312, + "grad_norm": 1.3984375, + "learning_rate": 0.00015688377026989626, + "loss": 0.28935771, + "memory(GiB)": 69.04, + "step": 11020, + "train_speed(iter/s)": 0.767361 + }, + { + "epoch": 1.0279329608938548, + "grad_norm": 1.5, + "learning_rate": 0.0001567148242349851, + "loss": 0.28245692, + "memory(GiB)": 31.11, + "step": 11040, + "train_speed(iter/s)": 0.767465 + }, + { + "epoch": 1.0297951582867784, + "grad_norm": 1.328125, + "learning_rate": 0.0001565456392220906, + "loss": 0.28632317, + "memory(GiB)": 31.11, + "step": 11060, + "train_speed(iter/s)": 0.767508 + }, + { + "epoch": 1.031657355679702, + "grad_norm": 1.484375, + "learning_rate": 0.00015637621594410383, + "loss": 0.30479004, + "memory(GiB)": 31.11, + "step": 11080, + "train_speed(iter/s)": 0.767594 + }, + { + "epoch": 1.0335195530726258, + "grad_norm": 1.453125, + "learning_rate": 0.00015620655511491984, + "loss": 0.27974844, + "memory(GiB)": 31.11, + "step": 11100, + "train_speed(iter/s)": 0.767652 + }, + { + "epoch": 1.0353817504655494, + "grad_norm": 1.5625, + "learning_rate": 0.00015603665744943473, + "loss": 0.26321175, + "memory(GiB)": 31.11, + "step": 11120, + "train_speed(iter/s)": 0.767723 + }, + { + "epoch": 1.037243947858473, + "grad_norm": 1.2578125, + "learning_rate": 0.0001558665236635424, + "loss": 0.27321138, + "memory(GiB)": 31.11, + "step": 11140, + "train_speed(iter/s)": 0.767764 + }, + { + "epoch": 1.0391061452513966, + "grad_norm": 1.765625, + "learning_rate": 0.00015569615447413184, + "loss": 0.28301878, + "memory(GiB)": 31.11, + "step": 11160, + "train_speed(iter/s)": 0.767868 + }, + { + "epoch": 1.0409683426443204, + "grad_norm": 1.2734375, + "learning_rate": 0.00015552555059908382, + "loss": 0.28938766, + "memory(GiB)": 31.11, + "step": 11180, + "train_speed(iter/s)": 0.767873 + }, + { + "epoch": 1.042830540037244, + "grad_norm": 1.4375, + "learning_rate": 0.00015535471275726813, + "loss": 0.27508519, + "memory(GiB)": 31.11, + "step": 11200, + "train_speed(iter/s)": 0.767984 + }, + { + "epoch": 1.0446927374301676, + "grad_norm": 1.546875, + "learning_rate": 0.0001551836416685403, + "loss": 0.27724741, + "memory(GiB)": 31.11, + "step": 11220, + "train_speed(iter/s)": 0.768059 + }, + { + "epoch": 1.0465549348230911, + "grad_norm": 1.390625, + "learning_rate": 0.00015501233805373876, + "loss": 0.26528058, + "memory(GiB)": 31.11, + "step": 11240, + "train_speed(iter/s)": 0.768095 + }, + { + "epoch": 1.048417132216015, + "grad_norm": 1.3515625, + "learning_rate": 0.0001548408026346817, + "loss": 0.28528447, + "memory(GiB)": 31.11, + "step": 11260, + "train_speed(iter/s)": 0.768155 + }, + { + "epoch": 1.0502793296089385, + "grad_norm": 1.875, + "learning_rate": 0.0001546690361341641, + "loss": 0.29369481, + "memory(GiB)": 31.11, + "step": 11280, + "train_speed(iter/s)": 0.768221 + }, + { + "epoch": 1.0521415270018621, + "grad_norm": 2.15625, + "learning_rate": 0.0001544970392759546, + "loss": 0.29468055, + "memory(GiB)": 31.11, + "step": 11300, + "train_speed(iter/s)": 0.768263 + }, + { + "epoch": 1.0540037243947857, + "grad_norm": 1.1796875, + "learning_rate": 0.0001543248127847925, + "loss": 0.28771296, + "memory(GiB)": 31.11, + "step": 11320, + "train_speed(iter/s)": 0.768313 + }, + { + "epoch": 1.0558659217877095, + "grad_norm": 1.515625, + "learning_rate": 0.00015415235738638468, + "loss": 0.30344753, + "memory(GiB)": 31.11, + "step": 11340, + "train_speed(iter/s)": 0.768354 + }, + { + "epoch": 1.0577281191806331, + "grad_norm": 1.4296875, + "learning_rate": 0.00015397967380740264, + "loss": 0.283692, + "memory(GiB)": 31.11, + "step": 11360, + "train_speed(iter/s)": 0.76837 + }, + { + "epoch": 1.0595903165735567, + "grad_norm": 1.609375, + "learning_rate": 0.00015380676277547927, + "loss": 0.270574, + "memory(GiB)": 31.11, + "step": 11380, + "train_speed(iter/s)": 0.768391 + }, + { + "epoch": 1.0614525139664805, + "grad_norm": 1.703125, + "learning_rate": 0.0001536336250192059, + "loss": 0.2845499, + "memory(GiB)": 31.11, + "step": 11400, + "train_speed(iter/s)": 0.768457 + }, + { + "epoch": 1.0633147113594041, + "grad_norm": 1.609375, + "learning_rate": 0.00015346026126812924, + "loss": 0.26141229, + "memory(GiB)": 31.11, + "step": 11420, + "train_speed(iter/s)": 0.768433 + }, + { + "epoch": 1.0651769087523277, + "grad_norm": 1.4453125, + "learning_rate": 0.0001532866722527482, + "loss": 0.28916895, + "memory(GiB)": 31.11, + "step": 11440, + "train_speed(iter/s)": 0.768458 + }, + { + "epoch": 1.0670391061452513, + "grad_norm": 1.3828125, + "learning_rate": 0.000153112858704511, + "loss": 0.29057674, + "memory(GiB)": 31.11, + "step": 11460, + "train_speed(iter/s)": 0.768521 + }, + { + "epoch": 1.0689013035381751, + "grad_norm": 1.5703125, + "learning_rate": 0.0001529388213558118, + "loss": 0.29088101, + "memory(GiB)": 31.11, + "step": 11480, + "train_speed(iter/s)": 0.76856 + }, + { + "epoch": 1.0707635009310987, + "grad_norm": 1.3125, + "learning_rate": 0.0001527645609399879, + "loss": 0.31489258, + "memory(GiB)": 31.11, + "step": 11500, + "train_speed(iter/s)": 0.768607 + }, + { + "epoch": 1.0726256983240223, + "grad_norm": 1.7265625, + "learning_rate": 0.00015259007819131657, + "loss": 0.29952948, + "memory(GiB)": 31.11, + "step": 11520, + "train_speed(iter/s)": 0.768637 + }, + { + "epoch": 1.074487895716946, + "grad_norm": 1.53125, + "learning_rate": 0.0001524153738450118, + "loss": 0.28699169, + "memory(GiB)": 31.11, + "step": 11540, + "train_speed(iter/s)": 0.76869 + }, + { + "epoch": 1.0763500931098697, + "grad_norm": 2.078125, + "learning_rate": 0.00015224044863722136, + "loss": 0.2873806, + "memory(GiB)": 31.11, + "step": 11560, + "train_speed(iter/s)": 0.768776 + }, + { + "epoch": 1.0782122905027933, + "grad_norm": 1.6484375, + "learning_rate": 0.0001520653033050237, + "loss": 0.26496642, + "memory(GiB)": 31.11, + "step": 11580, + "train_speed(iter/s)": 0.768813 + }, + { + "epoch": 1.080074487895717, + "grad_norm": 1.453125, + "learning_rate": 0.0001518899385864248, + "loss": 0.29042182, + "memory(GiB)": 31.11, + "step": 11600, + "train_speed(iter/s)": 0.768799 + }, + { + "epoch": 1.0819366852886405, + "grad_norm": 1.2265625, + "learning_rate": 0.000151714355220355, + "loss": 0.26191492, + "memory(GiB)": 31.11, + "step": 11620, + "train_speed(iter/s)": 0.768819 + }, + { + "epoch": 1.0837988826815643, + "grad_norm": 1.3671875, + "learning_rate": 0.00015153855394666604, + "loss": 0.29819686, + "memory(GiB)": 31.11, + "step": 11640, + "train_speed(iter/s)": 0.768815 + }, + { + "epoch": 1.085661080074488, + "grad_norm": 1.625, + "learning_rate": 0.0001513625355061278, + "loss": 0.26143956, + "memory(GiB)": 31.11, + "step": 11660, + "train_speed(iter/s)": 0.76891 + }, + { + "epoch": 1.0875232774674115, + "grad_norm": 1.546875, + "learning_rate": 0.0001511863006404252, + "loss": 0.2926599, + "memory(GiB)": 31.11, + "step": 11680, + "train_speed(iter/s)": 0.76896 + }, + { + "epoch": 1.089385474860335, + "grad_norm": 1.390625, + "learning_rate": 0.0001510098500921552, + "loss": 0.26912103, + "memory(GiB)": 31.11, + "step": 11700, + "train_speed(iter/s)": 0.768983 + }, + { + "epoch": 1.091247672253259, + "grad_norm": 1.6875, + "learning_rate": 0.00015083318460482344, + "loss": 0.2636065, + "memory(GiB)": 31.11, + "step": 11720, + "train_speed(iter/s)": 0.769047 + }, + { + "epoch": 1.0931098696461825, + "grad_norm": 1.1953125, + "learning_rate": 0.00015065630492284136, + "loss": 0.2775203, + "memory(GiB)": 31.11, + "step": 11740, + "train_speed(iter/s)": 0.769099 + }, + { + "epoch": 1.094972067039106, + "grad_norm": 1.6015625, + "learning_rate": 0.0001504792117915229, + "loss": 0.27552454, + "memory(GiB)": 31.11, + "step": 11760, + "train_speed(iter/s)": 0.769134 + }, + { + "epoch": 1.0968342644320297, + "grad_norm": 1.5234375, + "learning_rate": 0.0001503019059570815, + "loss": 0.23976431, + "memory(GiB)": 31.11, + "step": 11780, + "train_speed(iter/s)": 0.769196 + }, + { + "epoch": 1.0986964618249535, + "grad_norm": 1.265625, + "learning_rate": 0.00015012438816662665, + "loss": 0.2832912, + "memory(GiB)": 31.11, + "step": 11800, + "train_speed(iter/s)": 0.76928 + }, + { + "epoch": 1.100558659217877, + "grad_norm": 1.453125, + "learning_rate": 0.00014994665916816117, + "loss": 0.28068914, + "memory(GiB)": 31.11, + "step": 11820, + "train_speed(iter/s)": 0.769327 + }, + { + "epoch": 1.1024208566108007, + "grad_norm": 1.3046875, + "learning_rate": 0.00014976871971057775, + "loss": 0.26431036, + "memory(GiB)": 31.11, + "step": 11840, + "train_speed(iter/s)": 0.769318 + }, + { + "epoch": 1.1042830540037243, + "grad_norm": 1.546875, + "learning_rate": 0.0001495905705436559, + "loss": 0.28522236, + "memory(GiB)": 31.11, + "step": 11860, + "train_speed(iter/s)": 0.769396 + }, + { + "epoch": 1.106145251396648, + "grad_norm": 1.59375, + "learning_rate": 0.00014941221241805867, + "loss": 0.28218701, + "memory(GiB)": 31.11, + "step": 11880, + "train_speed(iter/s)": 0.76947 + }, + { + "epoch": 1.1080074487895717, + "grad_norm": 1.4765625, + "learning_rate": 0.00014923364608532982, + "loss": 0.26047482, + "memory(GiB)": 31.11, + "step": 11900, + "train_speed(iter/s)": 0.769479 + }, + { + "epoch": 1.1098696461824953, + "grad_norm": 1.5234375, + "learning_rate": 0.00014905487229789022, + "loss": 0.29261403, + "memory(GiB)": 31.11, + "step": 11920, + "train_speed(iter/s)": 0.769492 + }, + { + "epoch": 1.111731843575419, + "grad_norm": 1.2265625, + "learning_rate": 0.00014887589180903504, + "loss": 0.2624248, + "memory(GiB)": 31.11, + "step": 11940, + "train_speed(iter/s)": 0.769556 + }, + { + "epoch": 1.1135940409683427, + "grad_norm": 1.375, + "learning_rate": 0.0001486967053729303, + "loss": 0.25690923, + "memory(GiB)": 31.11, + "step": 11960, + "train_speed(iter/s)": 0.76963 + }, + { + "epoch": 1.1154562383612663, + "grad_norm": 1.390625, + "learning_rate": 0.00014851731374460988, + "loss": 0.28093395, + "memory(GiB)": 31.11, + "step": 11980, + "train_speed(iter/s)": 0.769697 + }, + { + "epoch": 1.1173184357541899, + "grad_norm": 1.2890625, + "learning_rate": 0.0001483377176799723, + "loss": 0.28581772, + "memory(GiB)": 31.11, + "step": 12000, + "train_speed(iter/s)": 0.769759 + }, + { + "epoch": 1.1173184357541899, + "eval_loss": 0.41399523615837097, + "eval_runtime": 77.0372, + "eval_samples_per_second": 180.237, + "eval_steps_per_second": 1.415, + "step": 12000 + }, + { + "epoch": 1.1191806331471137, + "grad_norm": 1.5703125, + "learning_rate": 0.0001481579179357774, + "loss": 0.28383641, + "memory(GiB)": 41.92, + "step": 12020, + "train_speed(iter/s)": 0.7643 + }, + { + "epoch": 1.1210428305400373, + "grad_norm": 1.6953125, + "learning_rate": 0.0001479779152696434, + "loss": 0.30731783, + "memory(GiB)": 41.92, + "step": 12040, + "train_speed(iter/s)": 0.764324 + }, + { + "epoch": 1.1229050279329609, + "grad_norm": 1.4296875, + "learning_rate": 0.00014779771044004346, + "loss": 0.26912603, + "memory(GiB)": 41.92, + "step": 12060, + "train_speed(iter/s)": 0.764388 + }, + { + "epoch": 1.1247672253258845, + "grad_norm": 1.5703125, + "learning_rate": 0.00014761730420630265, + "loss": 0.27448189, + "memory(GiB)": 41.92, + "step": 12080, + "train_speed(iter/s)": 0.764451 + }, + { + "epoch": 1.1266294227188083, + "grad_norm": 1.6953125, + "learning_rate": 0.00014743669732859464, + "loss": 0.28500652, + "memory(GiB)": 41.92, + "step": 12100, + "train_speed(iter/s)": 0.764504 + }, + { + "epoch": 1.1284916201117319, + "grad_norm": 1.140625, + "learning_rate": 0.0001472558905679386, + "loss": 0.26959038, + "memory(GiB)": 41.92, + "step": 12120, + "train_speed(iter/s)": 0.764615 + }, + { + "epoch": 1.1303538175046555, + "grad_norm": 1.390625, + "learning_rate": 0.00014707488468619596, + "loss": 0.28465414, + "memory(GiB)": 41.92, + "step": 12140, + "train_speed(iter/s)": 0.764627 + }, + { + "epoch": 1.132216014897579, + "grad_norm": 1.6328125, + "learning_rate": 0.00014689368044606708, + "loss": 0.2677213, + "memory(GiB)": 41.92, + "step": 12160, + "train_speed(iter/s)": 0.764649 + }, + { + "epoch": 1.1340782122905029, + "grad_norm": 1.6015625, + "learning_rate": 0.00014671227861108826, + "loss": 0.26625967, + "memory(GiB)": 41.92, + "step": 12180, + "train_speed(iter/s)": 0.764713 + }, + { + "epoch": 1.1359404096834265, + "grad_norm": 1.65625, + "learning_rate": 0.00014653067994562833, + "loss": 0.26140971, + "memory(GiB)": 41.92, + "step": 12200, + "train_speed(iter/s)": 0.764762 + }, + { + "epoch": 1.13780260707635, + "grad_norm": 1.265625, + "learning_rate": 0.00014634888521488547, + "loss": 0.27225456, + "memory(GiB)": 41.92, + "step": 12220, + "train_speed(iter/s)": 0.764812 + }, + { + "epoch": 1.1396648044692737, + "grad_norm": 1.484375, + "learning_rate": 0.00014616689518488416, + "loss": 0.27721281, + "memory(GiB)": 41.92, + "step": 12240, + "train_speed(iter/s)": 0.764902 + }, + { + "epoch": 1.1415270018621975, + "grad_norm": 1.1328125, + "learning_rate": 0.00014598471062247158, + "loss": 0.27392306, + "memory(GiB)": 41.92, + "step": 12260, + "train_speed(iter/s)": 0.764971 + }, + { + "epoch": 1.143389199255121, + "grad_norm": 1.46875, + "learning_rate": 0.00014580233229531482, + "loss": 0.26567361, + "memory(GiB)": 41.92, + "step": 12280, + "train_speed(iter/s)": 0.764996 + }, + { + "epoch": 1.1452513966480447, + "grad_norm": 1.859375, + "learning_rate": 0.00014561976097189736, + "loss": 0.25700495, + "memory(GiB)": 41.92, + "step": 12300, + "train_speed(iter/s)": 0.765049 + }, + { + "epoch": 1.1471135940409685, + "grad_norm": 1.9140625, + "learning_rate": 0.00014543699742151586, + "loss": 0.26121554, + "memory(GiB)": 41.92, + "step": 12320, + "train_speed(iter/s)": 0.76512 + }, + { + "epoch": 1.148975791433892, + "grad_norm": 1.5, + "learning_rate": 0.00014525404241427697, + "loss": 0.26690025, + "memory(GiB)": 41.92, + "step": 12340, + "train_speed(iter/s)": 0.765123 + }, + { + "epoch": 1.1508379888268156, + "grad_norm": 1.484375, + "learning_rate": 0.0001450708967210942, + "loss": 0.26092589, + "memory(GiB)": 41.92, + "step": 12360, + "train_speed(iter/s)": 0.765214 + }, + { + "epoch": 1.1527001862197392, + "grad_norm": 1.3046875, + "learning_rate": 0.0001448875611136843, + "loss": 0.24557509, + "memory(GiB)": 41.92, + "step": 12380, + "train_speed(iter/s)": 0.765292 + }, + { + "epoch": 1.1545623836126628, + "grad_norm": 1.3828125, + "learning_rate": 0.00014470403636456455, + "loss": 0.26972108, + "memory(GiB)": 41.92, + "step": 12400, + "train_speed(iter/s)": 0.765326 + }, + { + "epoch": 1.1564245810055866, + "grad_norm": 3.21875, + "learning_rate": 0.000144520323247049, + "loss": 0.26979203, + "memory(GiB)": 41.92, + "step": 12420, + "train_speed(iter/s)": 0.765415 + }, + { + "epoch": 1.1582867783985102, + "grad_norm": 1.3125, + "learning_rate": 0.0001443364225352455, + "loss": 0.26735349, + "memory(GiB)": 41.92, + "step": 12440, + "train_speed(iter/s)": 0.765454 + }, + { + "epoch": 1.1601489757914338, + "grad_norm": 1.78125, + "learning_rate": 0.00014415233500405237, + "loss": 0.2736985, + "memory(GiB)": 41.92, + "step": 12460, + "train_speed(iter/s)": 0.7655 + }, + { + "epoch": 1.1620111731843576, + "grad_norm": 1.859375, + "learning_rate": 0.0001439680614291551, + "loss": 0.27279396, + "memory(GiB)": 41.92, + "step": 12480, + "train_speed(iter/s)": 0.765512 + }, + { + "epoch": 1.1638733705772812, + "grad_norm": 1.53125, + "learning_rate": 0.00014378360258702311, + "loss": 0.2811708, + "memory(GiB)": 41.92, + "step": 12500, + "train_speed(iter/s)": 0.765564 + }, + { + "epoch": 1.1657355679702048, + "grad_norm": 1.3984375, + "learning_rate": 0.00014359895925490656, + "loss": 0.27203159, + "memory(GiB)": 41.92, + "step": 12520, + "train_speed(iter/s)": 0.765622 + }, + { + "epoch": 1.1675977653631284, + "grad_norm": 1.3671875, + "learning_rate": 0.00014341413221083282, + "loss": 0.25222397, + "memory(GiB)": 41.92, + "step": 12540, + "train_speed(iter/s)": 0.76566 + }, + { + "epoch": 1.169459962756052, + "grad_norm": 1.609375, + "learning_rate": 0.0001432291222336035, + "loss": 0.28008556, + "memory(GiB)": 41.92, + "step": 12560, + "train_speed(iter/s)": 0.765712 + }, + { + "epoch": 1.1713221601489758, + "grad_norm": 1.4921875, + "learning_rate": 0.00014304393010279104, + "loss": 0.27260323, + "memory(GiB)": 41.92, + "step": 12580, + "train_speed(iter/s)": 0.765741 + }, + { + "epoch": 1.1731843575418994, + "grad_norm": 1.1484375, + "learning_rate": 0.0001428585565987353, + "loss": 0.26112268, + "memory(GiB)": 41.92, + "step": 12600, + "train_speed(iter/s)": 0.765776 + }, + { + "epoch": 1.175046554934823, + "grad_norm": 1.171875, + "learning_rate": 0.0001426730025025405, + "loss": 0.27053833, + "memory(GiB)": 41.92, + "step": 12620, + "train_speed(iter/s)": 0.765859 + }, + { + "epoch": 1.1769087523277468, + "grad_norm": 1.296875, + "learning_rate": 0.00014248726859607174, + "loss": 0.25856524, + "memory(GiB)": 41.92, + "step": 12640, + "train_speed(iter/s)": 0.765898 + }, + { + "epoch": 1.1787709497206704, + "grad_norm": 1.3203125, + "learning_rate": 0.00014230135566195185, + "loss": 0.27097459, + "memory(GiB)": 41.92, + "step": 12660, + "train_speed(iter/s)": 0.765943 + }, + { + "epoch": 1.180633147113594, + "grad_norm": 1.5078125, + "learning_rate": 0.000142115264483558, + "loss": 0.24883418, + "memory(GiB)": 41.92, + "step": 12680, + "train_speed(iter/s)": 0.766016 + }, + { + "epoch": 1.1824953445065176, + "grad_norm": 1.96875, + "learning_rate": 0.00014192899584501841, + "loss": 0.27689738, + "memory(GiB)": 41.92, + "step": 12700, + "train_speed(iter/s)": 0.766087 + }, + { + "epoch": 1.1843575418994414, + "grad_norm": 1.59375, + "learning_rate": 0.00014174255053120904, + "loss": 0.26999578, + "memory(GiB)": 41.92, + "step": 12720, + "train_speed(iter/s)": 0.766176 + }, + { + "epoch": 1.186219739292365, + "grad_norm": 1.1484375, + "learning_rate": 0.00014155592932775033, + "loss": 0.25672078, + "memory(GiB)": 41.92, + "step": 12740, + "train_speed(iter/s)": 0.766239 + }, + { + "epoch": 1.1880819366852886, + "grad_norm": 1.6875, + "learning_rate": 0.0001413691330210039, + "loss": 0.24652848, + "memory(GiB)": 41.92, + "step": 12760, + "train_speed(iter/s)": 0.766343 + }, + { + "epoch": 1.1899441340782122, + "grad_norm": 1.46875, + "learning_rate": 0.0001411821623980691, + "loss": 0.26364939, + "memory(GiB)": 41.92, + "step": 12780, + "train_speed(iter/s)": 0.766431 + }, + { + "epoch": 1.191806331471136, + "grad_norm": 1.1640625, + "learning_rate": 0.00014099501824677992, + "loss": 0.27025316, + "memory(GiB)": 41.92, + "step": 12800, + "train_speed(iter/s)": 0.766497 + }, + { + "epoch": 1.1936685288640596, + "grad_norm": 1.28125, + "learning_rate": 0.00014080770135570142, + "loss": 0.22930434, + "memory(GiB)": 41.92, + "step": 12820, + "train_speed(iter/s)": 0.766501 + }, + { + "epoch": 1.1955307262569832, + "grad_norm": 1.3046875, + "learning_rate": 0.0001406202125141266, + "loss": 0.23245554, + "memory(GiB)": 41.92, + "step": 12840, + "train_speed(iter/s)": 0.766545 + }, + { + "epoch": 1.197392923649907, + "grad_norm": 1.5078125, + "learning_rate": 0.000140432552512073, + "loss": 0.25615866, + "memory(GiB)": 41.92, + "step": 12860, + "train_speed(iter/s)": 0.766559 + }, + { + "epoch": 1.1992551210428306, + "grad_norm": 1.296875, + "learning_rate": 0.00014024472214027939, + "loss": 0.24658422, + "memory(GiB)": 41.92, + "step": 12880, + "train_speed(iter/s)": 0.766583 + }, + { + "epoch": 1.2011173184357542, + "grad_norm": 1.46875, + "learning_rate": 0.00014005672219020234, + "loss": 0.24690557, + "memory(GiB)": 41.92, + "step": 12900, + "train_speed(iter/s)": 0.766626 + }, + { + "epoch": 1.2029795158286778, + "grad_norm": 1.6796875, + "learning_rate": 0.00013986855345401303, + "loss": 0.27605991, + "memory(GiB)": 41.92, + "step": 12920, + "train_speed(iter/s)": 0.76671 + }, + { + "epoch": 1.2048417132216014, + "grad_norm": 1.5703125, + "learning_rate": 0.00013968021672459388, + "loss": 0.24593453, + "memory(GiB)": 41.92, + "step": 12940, + "train_speed(iter/s)": 0.766726 + }, + { + "epoch": 1.2067039106145252, + "grad_norm": 1.390625, + "learning_rate": 0.00013949171279553516, + "loss": 0.2621069, + "memory(GiB)": 41.92, + "step": 12960, + "train_speed(iter/s)": 0.766796 + }, + { + "epoch": 1.2085661080074488, + "grad_norm": 1.71875, + "learning_rate": 0.00013930304246113158, + "loss": 0.26049728, + "memory(GiB)": 41.92, + "step": 12980, + "train_speed(iter/s)": 0.766867 + }, + { + "epoch": 1.2104283054003724, + "grad_norm": 1.4140625, + "learning_rate": 0.0001391142065163791, + "loss": 0.25473256, + "memory(GiB)": 41.92, + "step": 13000, + "train_speed(iter/s)": 0.76689 + }, + { + "epoch": 1.2122905027932962, + "grad_norm": 1.3671875, + "learning_rate": 0.00013892520575697163, + "loss": 0.25336313, + "memory(GiB)": 41.92, + "step": 13020, + "train_speed(iter/s)": 0.766867 + }, + { + "epoch": 1.2141527001862198, + "grad_norm": 1.359375, + "learning_rate": 0.0001387360409792973, + "loss": 0.26464946, + "memory(GiB)": 41.92, + "step": 13040, + "train_speed(iter/s)": 0.766959 + }, + { + "epoch": 1.2160148975791434, + "grad_norm": 1.453125, + "learning_rate": 0.00013854671298043557, + "loss": 0.26744711, + "memory(GiB)": 41.92, + "step": 13060, + "train_speed(iter/s)": 0.766995 + }, + { + "epoch": 1.217877094972067, + "grad_norm": 1.4765625, + "learning_rate": 0.0001383572225581535, + "loss": 0.25843668, + "memory(GiB)": 41.92, + "step": 13080, + "train_speed(iter/s)": 0.766992 + }, + { + "epoch": 1.2197392923649906, + "grad_norm": 1.3125, + "learning_rate": 0.00013816757051090273, + "loss": 0.24495487, + "memory(GiB)": 41.92, + "step": 13100, + "train_speed(iter/s)": 0.767065 + }, + { + "epoch": 1.2216014897579144, + "grad_norm": 1.453125, + "learning_rate": 0.00013797775763781578, + "loss": 0.24947252, + "memory(GiB)": 41.92, + "step": 13120, + "train_speed(iter/s)": 0.767141 + }, + { + "epoch": 1.223463687150838, + "grad_norm": 1.2265625, + "learning_rate": 0.0001377877847387029, + "loss": 0.24399962, + "memory(GiB)": 41.92, + "step": 13140, + "train_speed(iter/s)": 0.767172 + }, + { + "epoch": 1.2253258845437616, + "grad_norm": 1.359375, + "learning_rate": 0.00013759765261404865, + "loss": 0.2646559, + "memory(GiB)": 41.92, + "step": 13160, + "train_speed(iter/s)": 0.767248 + }, + { + "epoch": 1.2271880819366854, + "grad_norm": 1.5, + "learning_rate": 0.00013740736206500852, + "loss": 0.23537984, + "memory(GiB)": 41.92, + "step": 13180, + "train_speed(iter/s)": 0.767296 + }, + { + "epoch": 1.229050279329609, + "grad_norm": 1.8359375, + "learning_rate": 0.0001372169138934055, + "loss": 0.2611448, + "memory(GiB)": 41.92, + "step": 13200, + "train_speed(iter/s)": 0.767357 + }, + { + "epoch": 1.2309124767225326, + "grad_norm": 1.3671875, + "learning_rate": 0.00013702630890172681, + "loss": 0.24709506, + "memory(GiB)": 41.92, + "step": 13220, + "train_speed(iter/s)": 0.767381 + }, + { + "epoch": 1.2327746741154562, + "grad_norm": 1.7421875, + "learning_rate": 0.00013683554789312039, + "loss": 0.25667541, + "memory(GiB)": 41.92, + "step": 13240, + "train_speed(iter/s)": 0.767469 + }, + { + "epoch": 1.23463687150838, + "grad_norm": 1.203125, + "learning_rate": 0.00013664463167139165, + "loss": 0.25362358, + "memory(GiB)": 41.92, + "step": 13260, + "train_speed(iter/s)": 0.767528 + }, + { + "epoch": 1.2364990689013036, + "grad_norm": 1.3671875, + "learning_rate": 0.000136453561041, + "loss": 0.27490563, + "memory(GiB)": 41.92, + "step": 13280, + "train_speed(iter/s)": 0.767627 + }, + { + "epoch": 1.2383612662942272, + "grad_norm": 1.3984375, + "learning_rate": 0.00013626233680705548, + "loss": 0.2458004, + "memory(GiB)": 41.92, + "step": 13300, + "train_speed(iter/s)": 0.767693 + }, + { + "epoch": 1.2402234636871508, + "grad_norm": 1.375, + "learning_rate": 0.0001360709597753153, + "loss": 0.25720315, + "memory(GiB)": 41.92, + "step": 13320, + "train_speed(iter/s)": 0.767713 + }, + { + "epoch": 1.2420856610800746, + "grad_norm": 1.21875, + "learning_rate": 0.0001358794307521806, + "loss": 0.22600164, + "memory(GiB)": 41.92, + "step": 13340, + "train_speed(iter/s)": 0.767729 + }, + { + "epoch": 1.2439478584729982, + "grad_norm": 1.5546875, + "learning_rate": 0.000135687750544693, + "loss": 0.24354362, + "memory(GiB)": 41.92, + "step": 13360, + "train_speed(iter/s)": 0.767783 + }, + { + "epoch": 1.2458100558659218, + "grad_norm": 1.3125, + "learning_rate": 0.000135495919960531, + "loss": 0.24208927, + "memory(GiB)": 41.92, + "step": 13380, + "train_speed(iter/s)": 0.767773 + }, + { + "epoch": 1.2476722532588453, + "grad_norm": 1.453125, + "learning_rate": 0.00013530393980800686, + "loss": 0.24553039, + "memory(GiB)": 41.92, + "step": 13400, + "train_speed(iter/s)": 0.767854 + }, + { + "epoch": 1.2495344506517692, + "grad_norm": 1.234375, + "learning_rate": 0.00013511181089606304, + "loss": 0.23603408, + "memory(GiB)": 41.92, + "step": 13420, + "train_speed(iter/s)": 0.767905 + }, + { + "epoch": 1.2513966480446927, + "grad_norm": 1.2734375, + "learning_rate": 0.00013491953403426887, + "loss": 0.25591078, + "memory(GiB)": 41.92, + "step": 13440, + "train_speed(iter/s)": 0.767918 + }, + { + "epoch": 1.2532588454376163, + "grad_norm": 1.6171875, + "learning_rate": 0.00013472711003281696, + "loss": 0.23547509, + "memory(GiB)": 41.92, + "step": 13460, + "train_speed(iter/s)": 0.767974 + }, + { + "epoch": 1.25512104283054, + "grad_norm": 1.296875, + "learning_rate": 0.00013453453970252013, + "loss": 0.25424879, + "memory(GiB)": 41.92, + "step": 13480, + "train_speed(iter/s)": 0.768055 + }, + { + "epoch": 1.2569832402234637, + "grad_norm": 1.5546875, + "learning_rate": 0.00013434182385480757, + "loss": 0.24909823, + "memory(GiB)": 41.92, + "step": 13500, + "train_speed(iter/s)": 0.768135 + }, + { + "epoch": 1.2588454376163873, + "grad_norm": 1.2734375, + "learning_rate": 0.00013414896330172176, + "loss": 0.24665921, + "memory(GiB)": 41.92, + "step": 13520, + "train_speed(iter/s)": 0.76815 + }, + { + "epoch": 1.260707635009311, + "grad_norm": 1.421875, + "learning_rate": 0.0001339559588559149, + "loss": 0.25871291, + "memory(GiB)": 41.92, + "step": 13540, + "train_speed(iter/s)": 0.768194 + }, + { + "epoch": 1.2625698324022347, + "grad_norm": 1.1171875, + "learning_rate": 0.0001337628113306455, + "loss": 0.25705509, + "memory(GiB)": 41.92, + "step": 13560, + "train_speed(iter/s)": 0.768234 + }, + { + "epoch": 1.2644320297951583, + "grad_norm": 1.296875, + "learning_rate": 0.00013356952153977488, + "loss": 0.22989726, + "memory(GiB)": 41.92, + "step": 13580, + "train_speed(iter/s)": 0.768319 + }, + { + "epoch": 1.266294227188082, + "grad_norm": 1.546875, + "learning_rate": 0.00013337609029776402, + "loss": 0.25790679, + "memory(GiB)": 41.92, + "step": 13600, + "train_speed(iter/s)": 0.76841 + }, + { + "epoch": 1.2681564245810055, + "grad_norm": 1.296875, + "learning_rate": 0.00013318251841966974, + "loss": 0.24724863, + "memory(GiB)": 41.92, + "step": 13620, + "train_speed(iter/s)": 0.768461 + }, + { + "epoch": 1.2700186219739291, + "grad_norm": 1.40625, + "learning_rate": 0.0001329888067211415, + "loss": 0.25862865, + "memory(GiB)": 41.92, + "step": 13640, + "train_speed(iter/s)": 0.768508 + }, + { + "epoch": 1.271880819366853, + "grad_norm": 1.1875, + "learning_rate": 0.00013279495601841798, + "loss": 0.24088619, + "memory(GiB)": 41.92, + "step": 13660, + "train_speed(iter/s)": 0.768585 + }, + { + "epoch": 1.2737430167597765, + "grad_norm": 1.421875, + "learning_rate": 0.00013260096712832355, + "loss": 0.25227604, + "memory(GiB)": 41.92, + "step": 13680, + "train_speed(iter/s)": 0.768636 + }, + { + "epoch": 1.2756052141527001, + "grad_norm": 1.203125, + "learning_rate": 0.00013240684086826483, + "loss": 0.24656982, + "memory(GiB)": 41.92, + "step": 13700, + "train_speed(iter/s)": 0.768727 + }, + { + "epoch": 1.277467411545624, + "grad_norm": 1.1953125, + "learning_rate": 0.00013221257805622728, + "loss": 0.25998006, + "memory(GiB)": 41.92, + "step": 13720, + "train_speed(iter/s)": 0.768793 + }, + { + "epoch": 1.2793296089385475, + "grad_norm": 1.3203125, + "learning_rate": 0.00013201817951077175, + "loss": 0.24141338, + "memory(GiB)": 41.92, + "step": 13740, + "train_speed(iter/s)": 0.768859 + }, + { + "epoch": 1.2811918063314711, + "grad_norm": 1.375, + "learning_rate": 0.00013182364605103107, + "loss": 0.22881098, + "memory(GiB)": 41.92, + "step": 13760, + "train_speed(iter/s)": 0.768936 + }, + { + "epoch": 1.2830540037243947, + "grad_norm": 1.6015625, + "learning_rate": 0.0001316289784967065, + "loss": 0.21602743, + "memory(GiB)": 41.92, + "step": 13780, + "train_speed(iter/s)": 0.769004 + }, + { + "epoch": 1.2849162011173183, + "grad_norm": 1.421875, + "learning_rate": 0.0001314341776680643, + "loss": 0.23749352, + "memory(GiB)": 41.92, + "step": 13800, + "train_speed(iter/s)": 0.769016 + }, + { + "epoch": 1.2867783985102421, + "grad_norm": 1.6171875, + "learning_rate": 0.00013123924438593245, + "loss": 0.23468378, + "memory(GiB)": 41.92, + "step": 13820, + "train_speed(iter/s)": 0.769095 + }, + { + "epoch": 1.2886405959031657, + "grad_norm": 1.1484375, + "learning_rate": 0.0001310441794716969, + "loss": 0.23531232, + "memory(GiB)": 58.24, + "step": 13840, + "train_speed(iter/s)": 0.769107 + }, + { + "epoch": 1.2905027932960893, + "grad_norm": 1.375, + "learning_rate": 0.00013084898374729825, + "loss": 0.25231352, + "memory(GiB)": 58.24, + "step": 13860, + "train_speed(iter/s)": 0.769127 + }, + { + "epoch": 1.2923649906890131, + "grad_norm": 1.3984375, + "learning_rate": 0.00013065365803522846, + "loss": 0.22567735, + "memory(GiB)": 58.24, + "step": 13880, + "train_speed(iter/s)": 0.769166 + }, + { + "epoch": 1.2942271880819367, + "grad_norm": 1.5625, + "learning_rate": 0.00013045820315852702, + "loss": 0.21936474, + "memory(GiB)": 58.24, + "step": 13900, + "train_speed(iter/s)": 0.769171 + }, + { + "epoch": 1.2960893854748603, + "grad_norm": 1.265625, + "learning_rate": 0.0001302626199407778, + "loss": 0.24280531, + "memory(GiB)": 58.24, + "step": 13920, + "train_speed(iter/s)": 0.769216 + }, + { + "epoch": 1.2979515828677841, + "grad_norm": 1.4609375, + "learning_rate": 0.0001300669092061054, + "loss": 0.23217328, + "memory(GiB)": 58.24, + "step": 13940, + "train_speed(iter/s)": 0.769238 + }, + { + "epoch": 1.2998137802607077, + "grad_norm": 1.5703125, + "learning_rate": 0.00012987107177917176, + "loss": 0.22496073, + "memory(GiB)": 58.24, + "step": 13960, + "train_speed(iter/s)": 0.769301 + }, + { + "epoch": 1.3016759776536313, + "grad_norm": 1.609375, + "learning_rate": 0.00012967510848517266, + "loss": 0.23814521, + "memory(GiB)": 58.24, + "step": 13980, + "train_speed(iter/s)": 0.769345 + }, + { + "epoch": 1.303538175046555, + "grad_norm": 1.1875, + "learning_rate": 0.00012947902014983425, + "loss": 0.22658379, + "memory(GiB)": 58.24, + "step": 14000, + "train_speed(iter/s)": 0.769362 + }, + { + "epoch": 1.303538175046555, + "eval_loss": 0.40084439516067505, + "eval_runtime": 77.0745, + "eval_samples_per_second": 180.15, + "eval_steps_per_second": 1.414, + "step": 14000 + }, + { + "epoch": 1.3054003724394785, + "grad_norm": 1.515625, + "learning_rate": 0.0001292828075994095, + "loss": 0.21561022, + "memory(GiB)": 58.24, + "step": 14020, + "train_speed(iter/s)": 0.764795 + }, + { + "epoch": 1.3072625698324023, + "grad_norm": 1.2421875, + "learning_rate": 0.00012908647166067494, + "loss": 0.22430439, + "memory(GiB)": 58.24, + "step": 14040, + "train_speed(iter/s)": 0.764864 + }, + { + "epoch": 1.309124767225326, + "grad_norm": 1.8203125, + "learning_rate": 0.00012889001316092682, + "loss": 0.22065244, + "memory(GiB)": 58.24, + "step": 14060, + "train_speed(iter/s)": 0.76492 + }, + { + "epoch": 1.3109869646182495, + "grad_norm": 1.0859375, + "learning_rate": 0.00012869343292797795, + "loss": 0.24164536, + "memory(GiB)": 58.24, + "step": 14080, + "train_speed(iter/s)": 0.764972 + }, + { + "epoch": 1.3128491620111733, + "grad_norm": 1.375, + "learning_rate": 0.00012849673179015402, + "loss": 0.2178977, + "memory(GiB)": 58.24, + "step": 14100, + "train_speed(iter/s)": 0.764989 + }, + { + "epoch": 1.314711359404097, + "grad_norm": 1.390625, + "learning_rate": 0.00012829991057629022, + "loss": 0.24479382, + "memory(GiB)": 58.24, + "step": 14120, + "train_speed(iter/s)": 0.765055 + }, + { + "epoch": 1.3165735567970205, + "grad_norm": 1.578125, + "learning_rate": 0.00012810297011572771, + "loss": 0.24034092, + "memory(GiB)": 58.24, + "step": 14140, + "train_speed(iter/s)": 0.765145 + }, + { + "epoch": 1.318435754189944, + "grad_norm": 1.546875, + "learning_rate": 0.0001279059112383101, + "loss": 0.2230279, + "memory(GiB)": 58.24, + "step": 14160, + "train_speed(iter/s)": 0.765184 + }, + { + "epoch": 1.3202979515828677, + "grad_norm": 1.515625, + "learning_rate": 0.00012770873477437994, + "loss": 0.23335476, + "memory(GiB)": 58.24, + "step": 14180, + "train_speed(iter/s)": 0.765219 + }, + { + "epoch": 1.3221601489757915, + "grad_norm": 1.5390625, + "learning_rate": 0.00012751144155477527, + "loss": 0.23615379, + "memory(GiB)": 58.24, + "step": 14200, + "train_speed(iter/s)": 0.765242 + }, + { + "epoch": 1.324022346368715, + "grad_norm": 1.25, + "learning_rate": 0.0001273140324108261, + "loss": 0.22382288, + "memory(GiB)": 58.24, + "step": 14220, + "train_speed(iter/s)": 0.76528 + }, + { + "epoch": 1.3258845437616387, + "grad_norm": 1.8828125, + "learning_rate": 0.00012711650817435091, + "loss": 0.24886432, + "memory(GiB)": 58.24, + "step": 14240, + "train_speed(iter/s)": 0.765335 + }, + { + "epoch": 1.3277467411545625, + "grad_norm": 1.28125, + "learning_rate": 0.0001269188696776532, + "loss": 0.25288081, + "memory(GiB)": 58.24, + "step": 14260, + "train_speed(iter/s)": 0.765389 + }, + { + "epoch": 1.329608938547486, + "grad_norm": 1.296875, + "learning_rate": 0.00012672111775351781, + "loss": 0.23030953, + "memory(GiB)": 58.24, + "step": 14280, + "train_speed(iter/s)": 0.765397 + }, + { + "epoch": 1.3314711359404097, + "grad_norm": 1.296875, + "learning_rate": 0.00012652325323520762, + "loss": 0.23378878, + "memory(GiB)": 58.24, + "step": 14300, + "train_speed(iter/s)": 0.765441 + }, + { + "epoch": 1.3333333333333333, + "grad_norm": 1.3203125, + "learning_rate": 0.00012632527695645993, + "loss": 0.22340455, + "memory(GiB)": 58.24, + "step": 14320, + "train_speed(iter/s)": 0.765486 + }, + { + "epoch": 1.3351955307262569, + "grad_norm": 1.296875, + "learning_rate": 0.0001261271897514829, + "loss": 0.22818685, + "memory(GiB)": 58.24, + "step": 14340, + "train_speed(iter/s)": 0.765531 + }, + { + "epoch": 1.3370577281191807, + "grad_norm": 1.1953125, + "learning_rate": 0.0001259289924549522, + "loss": 0.22412224, + "memory(GiB)": 58.24, + "step": 14360, + "train_speed(iter/s)": 0.765583 + }, + { + "epoch": 1.3389199255121043, + "grad_norm": 1.3984375, + "learning_rate": 0.00012573068590200731, + "loss": 0.23063211, + "memory(GiB)": 58.24, + "step": 14380, + "train_speed(iter/s)": 0.765598 + }, + { + "epoch": 1.3407821229050279, + "grad_norm": 1.5, + "learning_rate": 0.0001255322709282481, + "loss": 0.21774821, + "memory(GiB)": 58.24, + "step": 14400, + "train_speed(iter/s)": 0.765623 + }, + { + "epoch": 1.3426443202979517, + "grad_norm": 1.2734375, + "learning_rate": 0.0001253337483697313, + "loss": 0.22045803, + "memory(GiB)": 58.24, + "step": 14420, + "train_speed(iter/s)": 0.76571 + }, + { + "epoch": 1.3445065176908753, + "grad_norm": 1.671875, + "learning_rate": 0.00012513511906296696, + "loss": 0.23508878, + "memory(GiB)": 58.24, + "step": 14440, + "train_speed(iter/s)": 0.765764 + }, + { + "epoch": 1.3463687150837989, + "grad_norm": 1.171875, + "learning_rate": 0.00012493638384491492, + "loss": 0.20511746, + "memory(GiB)": 58.24, + "step": 14460, + "train_speed(iter/s)": 0.765794 + }, + { + "epoch": 1.3482309124767227, + "grad_norm": 1.078125, + "learning_rate": 0.00012473754355298136, + "loss": 0.24029357, + "memory(GiB)": 58.24, + "step": 14480, + "train_speed(iter/s)": 0.76582 + }, + { + "epoch": 1.3500931098696463, + "grad_norm": 1.1328125, + "learning_rate": 0.00012453859902501516, + "loss": 0.23441067, + "memory(GiB)": 58.24, + "step": 14500, + "train_speed(iter/s)": 0.765812 + }, + { + "epoch": 1.3519553072625698, + "grad_norm": 1.3671875, + "learning_rate": 0.00012433955109930432, + "loss": 0.2113512, + "memory(GiB)": 58.24, + "step": 14520, + "train_speed(iter/s)": 0.76589 + }, + { + "epoch": 1.3538175046554934, + "grad_norm": 1.53125, + "learning_rate": 0.00012414040061457275, + "loss": 0.21798153, + "memory(GiB)": 58.24, + "step": 14540, + "train_speed(iter/s)": 0.765908 + }, + { + "epoch": 1.355679702048417, + "grad_norm": 1.140625, + "learning_rate": 0.00012394114840997634, + "loss": 0.21171179, + "memory(GiB)": 58.24, + "step": 14560, + "train_speed(iter/s)": 0.765942 + }, + { + "epoch": 1.3575418994413408, + "grad_norm": 1.0546875, + "learning_rate": 0.00012374179532509958, + "loss": 0.2068325, + "memory(GiB)": 58.24, + "step": 14580, + "train_speed(iter/s)": 0.765988 + }, + { + "epoch": 1.3594040968342644, + "grad_norm": 1.140625, + "learning_rate": 0.00012354234219995214, + "loss": 0.21365037, + "memory(GiB)": 58.24, + "step": 14600, + "train_speed(iter/s)": 0.766031 + }, + { + "epoch": 1.361266294227188, + "grad_norm": 1.3359375, + "learning_rate": 0.00012334278987496518, + "loss": 0.21723585, + "memory(GiB)": 58.24, + "step": 14620, + "train_speed(iter/s)": 0.766081 + }, + { + "epoch": 1.3631284916201118, + "grad_norm": 1.1640625, + "learning_rate": 0.00012314313919098787, + "loss": 0.21648164, + "memory(GiB)": 58.24, + "step": 14640, + "train_speed(iter/s)": 0.76614 + }, + { + "epoch": 1.3649906890130354, + "grad_norm": 1.1328125, + "learning_rate": 0.00012294339098928381, + "loss": 0.22302213, + "memory(GiB)": 58.24, + "step": 14660, + "train_speed(iter/s)": 0.766204 + }, + { + "epoch": 1.366852886405959, + "grad_norm": 1.4765625, + "learning_rate": 0.00012274354611152753, + "loss": 0.22709019, + "memory(GiB)": 58.24, + "step": 14680, + "train_speed(iter/s)": 0.766276 + }, + { + "epoch": 1.3687150837988826, + "grad_norm": 1.484375, + "learning_rate": 0.00012254360539980087, + "loss": 0.21800654, + "memory(GiB)": 58.24, + "step": 14700, + "train_speed(iter/s)": 0.76635 + }, + { + "epoch": 1.3705772811918062, + "grad_norm": 1.3515625, + "learning_rate": 0.0001223435696965896, + "loss": 0.22402124, + "memory(GiB)": 58.24, + "step": 14720, + "train_speed(iter/s)": 0.766429 + }, + { + "epoch": 1.37243947858473, + "grad_norm": 1.3046875, + "learning_rate": 0.0001221434398447796, + "loss": 0.210607, + "memory(GiB)": 58.24, + "step": 14740, + "train_speed(iter/s)": 0.766452 + }, + { + "epoch": 1.3743016759776536, + "grad_norm": 1.4296875, + "learning_rate": 0.00012194321668765356, + "loss": 0.2248714, + "memory(GiB)": 58.24, + "step": 14760, + "train_speed(iter/s)": 0.766523 + }, + { + "epoch": 1.3761638733705772, + "grad_norm": 1.1796875, + "learning_rate": 0.00012174290106888731, + "loss": 0.21417339, + "memory(GiB)": 58.24, + "step": 14780, + "train_speed(iter/s)": 0.766543 + }, + { + "epoch": 1.378026070763501, + "grad_norm": 1.546875, + "learning_rate": 0.00012154249383254632, + "loss": 0.234969, + "memory(GiB)": 58.24, + "step": 14800, + "train_speed(iter/s)": 0.76655 + }, + { + "epoch": 1.3798882681564246, + "grad_norm": 1.1640625, + "learning_rate": 0.00012134199582308204, + "loss": 0.21318011, + "memory(GiB)": 58.24, + "step": 14820, + "train_speed(iter/s)": 0.766531 + }, + { + "epoch": 1.3817504655493482, + "grad_norm": 1.265625, + "learning_rate": 0.00012114140788532842, + "loss": 0.22573395, + "memory(GiB)": 58.24, + "step": 14840, + "train_speed(iter/s)": 0.766585 + }, + { + "epoch": 1.3836126629422718, + "grad_norm": 1.4140625, + "learning_rate": 0.00012094073086449837, + "loss": 0.21785412, + "memory(GiB)": 58.24, + "step": 14860, + "train_speed(iter/s)": 0.766626 + }, + { + "epoch": 1.3854748603351954, + "grad_norm": 1.375, + "learning_rate": 0.00012073996560618019, + "loss": 0.24774947, + "memory(GiB)": 58.24, + "step": 14880, + "train_speed(iter/s)": 0.766689 + }, + { + "epoch": 1.3873370577281192, + "grad_norm": 1.4453125, + "learning_rate": 0.0001205391129563339, + "loss": 0.23350339, + "memory(GiB)": 58.24, + "step": 14900, + "train_speed(iter/s)": 0.766772 + }, + { + "epoch": 1.3891992551210428, + "grad_norm": 1.3828125, + "learning_rate": 0.0001203381737612878, + "loss": 0.19333084, + "memory(GiB)": 58.24, + "step": 14920, + "train_speed(iter/s)": 0.766812 + }, + { + "epoch": 1.3910614525139664, + "grad_norm": 1.1171875, + "learning_rate": 0.0001201371488677349, + "loss": 0.21416593, + "memory(GiB)": 58.24, + "step": 14940, + "train_speed(iter/s)": 0.766851 + }, + { + "epoch": 1.3929236499068902, + "grad_norm": 1.4296875, + "learning_rate": 0.0001199360391227293, + "loss": 0.21936598, + "memory(GiB)": 58.24, + "step": 14960, + "train_speed(iter/s)": 0.766868 + }, + { + "epoch": 1.3947858472998138, + "grad_norm": 0.94140625, + "learning_rate": 0.0001197348453736826, + "loss": 0.22091618, + "memory(GiB)": 58.24, + "step": 14980, + "train_speed(iter/s)": 0.766953 + }, + { + "epoch": 1.3966480446927374, + "grad_norm": 1.328125, + "learning_rate": 0.00011953356846836037, + "loss": 0.21947927, + "memory(GiB)": 58.24, + "step": 15000, + "train_speed(iter/s)": 0.767002 + }, + { + "epoch": 1.3985102420856612, + "grad_norm": 1.203125, + "learning_rate": 0.00011933220925487862, + "loss": 0.23145449, + "memory(GiB)": 58.24, + "step": 15020, + "train_speed(iter/s)": 0.767081 + }, + { + "epoch": 1.4003724394785848, + "grad_norm": 1.296875, + "learning_rate": 0.0001191307685817002, + "loss": 0.22147784, + "memory(GiB)": 58.24, + "step": 15040, + "train_speed(iter/s)": 0.767163 + }, + { + "epoch": 1.4022346368715084, + "grad_norm": 1.2734375, + "learning_rate": 0.00011892924729763109, + "loss": 0.19200894, + "memory(GiB)": 58.24, + "step": 15060, + "train_speed(iter/s)": 0.767203 + }, + { + "epoch": 1.404096834264432, + "grad_norm": 1.484375, + "learning_rate": 0.00011872764625181701, + "loss": 0.21943903, + "memory(GiB)": 58.24, + "step": 15080, + "train_speed(iter/s)": 0.767267 + }, + { + "epoch": 1.4059590316573556, + "grad_norm": 1.125, + "learning_rate": 0.00011852596629373977, + "loss": 0.21660228, + "memory(GiB)": 58.24, + "step": 15100, + "train_speed(iter/s)": 0.767307 + }, + { + "epoch": 1.4078212290502794, + "grad_norm": 0.94921875, + "learning_rate": 0.00011832420827321373, + "loss": 0.21862273, + "memory(GiB)": 58.24, + "step": 15120, + "train_speed(iter/s)": 0.767327 + }, + { + "epoch": 1.409683426443203, + "grad_norm": 1.3125, + "learning_rate": 0.00011812237304038209, + "loss": 0.24118257, + "memory(GiB)": 58.24, + "step": 15140, + "train_speed(iter/s)": 0.767389 + }, + { + "epoch": 1.4115456238361266, + "grad_norm": 1.40625, + "learning_rate": 0.00011792046144571348, + "loss": 0.20173512, + "memory(GiB)": 58.24, + "step": 15160, + "train_speed(iter/s)": 0.767455 + }, + { + "epoch": 1.4134078212290504, + "grad_norm": 1.2734375, + "learning_rate": 0.00011771847433999823, + "loss": 0.19333193, + "memory(GiB)": 58.24, + "step": 15180, + "train_speed(iter/s)": 0.767496 + }, + { + "epoch": 1.415270018621974, + "grad_norm": 1.5390625, + "learning_rate": 0.00011751641257434492, + "loss": 0.21858253, + "memory(GiB)": 58.24, + "step": 15200, + "train_speed(iter/s)": 0.767558 + }, + { + "epoch": 1.4171322160148976, + "grad_norm": 1.2734375, + "learning_rate": 0.00011731427700017663, + "loss": 0.20099108, + "memory(GiB)": 58.24, + "step": 15220, + "train_speed(iter/s)": 0.76761 + }, + { + "epoch": 1.4189944134078212, + "grad_norm": 1.265625, + "learning_rate": 0.00011711206846922753, + "loss": 0.20006487, + "memory(GiB)": 58.24, + "step": 15240, + "train_speed(iter/s)": 0.767685 + }, + { + "epoch": 1.4208566108007448, + "grad_norm": 1.265625, + "learning_rate": 0.00011690978783353914, + "loss": 0.20852513, + "memory(GiB)": 58.24, + "step": 15260, + "train_speed(iter/s)": 0.767757 + }, + { + "epoch": 1.4227188081936686, + "grad_norm": 1.015625, + "learning_rate": 0.00011670743594545692, + "loss": 0.19811587, + "memory(GiB)": 58.24, + "step": 15280, + "train_speed(iter/s)": 0.767769 + }, + { + "epoch": 1.4245810055865922, + "grad_norm": 1.046875, + "learning_rate": 0.00011650501365762638, + "loss": 0.23192661, + "memory(GiB)": 58.24, + "step": 15300, + "train_speed(iter/s)": 0.76783 + }, + { + "epoch": 1.4264432029795158, + "grad_norm": 1.40625, + "learning_rate": 0.00011630252182298985, + "loss": 0.20126216, + "memory(GiB)": 58.24, + "step": 15320, + "train_speed(iter/s)": 0.767878 + }, + { + "epoch": 1.4283054003724396, + "grad_norm": 1.25, + "learning_rate": 0.00011609996129478257, + "loss": 0.1996449, + "memory(GiB)": 58.24, + "step": 15340, + "train_speed(iter/s)": 0.767869 + }, + { + "epoch": 1.4301675977653632, + "grad_norm": 1.375, + "learning_rate": 0.00011589733292652938, + "loss": 0.18534939, + "memory(GiB)": 58.24, + "step": 15360, + "train_speed(iter/s)": 0.767893 + }, + { + "epoch": 1.4320297951582868, + "grad_norm": 1.109375, + "learning_rate": 0.00011569463757204084, + "loss": 0.21407509, + "memory(GiB)": 58.24, + "step": 15380, + "train_speed(iter/s)": 0.76794 + }, + { + "epoch": 1.4338919925512104, + "grad_norm": 1.3046875, + "learning_rate": 0.00011549187608540983, + "loss": 0.18790734, + "memory(GiB)": 58.24, + "step": 15400, + "train_speed(iter/s)": 0.767948 + }, + { + "epoch": 1.435754189944134, + "grad_norm": 1.265625, + "learning_rate": 0.00011528904932100786, + "loss": 0.21195495, + "memory(GiB)": 58.24, + "step": 15420, + "train_speed(iter/s)": 0.768007 + }, + { + "epoch": 1.4376163873370578, + "grad_norm": 1.1015625, + "learning_rate": 0.00011508615813348156, + "loss": 0.22268994, + "memory(GiB)": 58.24, + "step": 15440, + "train_speed(iter/s)": 0.768083 + }, + { + "epoch": 1.4394785847299814, + "grad_norm": 1.4453125, + "learning_rate": 0.00011488320337774893, + "loss": 0.21396356, + "memory(GiB)": 58.24, + "step": 15460, + "train_speed(iter/s)": 0.768135 + }, + { + "epoch": 1.441340782122905, + "grad_norm": 1.1796875, + "learning_rate": 0.00011468018590899591, + "loss": 0.18995935, + "memory(GiB)": 58.24, + "step": 15480, + "train_speed(iter/s)": 0.768203 + }, + { + "epoch": 1.4432029795158288, + "grad_norm": 1.203125, + "learning_rate": 0.00011447710658267266, + "loss": 0.2040406, + "memory(GiB)": 58.24, + "step": 15500, + "train_speed(iter/s)": 0.768246 + }, + { + "epoch": 1.4450651769087524, + "grad_norm": 1.421875, + "learning_rate": 0.00011427396625448999, + "loss": 0.19963667, + "memory(GiB)": 58.24, + "step": 15520, + "train_speed(iter/s)": 0.768292 + }, + { + "epoch": 1.446927374301676, + "grad_norm": 1.421875, + "learning_rate": 0.00011407076578041569, + "loss": 0.20383754, + "memory(GiB)": 58.24, + "step": 15540, + "train_speed(iter/s)": 0.768307 + }, + { + "epoch": 1.4487895716945998, + "grad_norm": 1.3359375, + "learning_rate": 0.00011386750601667111, + "loss": 0.20976281, + "memory(GiB)": 58.24, + "step": 15560, + "train_speed(iter/s)": 0.768352 + }, + { + "epoch": 1.4506517690875234, + "grad_norm": 1.3359375, + "learning_rate": 0.00011366418781972732, + "loss": 0.21814973, + "memory(GiB)": 58.24, + "step": 15580, + "train_speed(iter/s)": 0.768338 + }, + { + "epoch": 1.452513966480447, + "grad_norm": 1.5078125, + "learning_rate": 0.00011346081204630166, + "loss": 0.20122726, + "memory(GiB)": 58.24, + "step": 15600, + "train_speed(iter/s)": 0.768344 + }, + { + "epoch": 1.4543761638733705, + "grad_norm": 1.03125, + "learning_rate": 0.00011325737955335405, + "loss": 0.20199561, + "memory(GiB)": 58.24, + "step": 15620, + "train_speed(iter/s)": 0.768395 + }, + { + "epoch": 1.4562383612662941, + "grad_norm": 1.203125, + "learning_rate": 0.00011305389119808345, + "loss": 0.19181294, + "memory(GiB)": 58.24, + "step": 15640, + "train_speed(iter/s)": 0.76846 + }, + { + "epoch": 1.458100558659218, + "grad_norm": 1.5078125, + "learning_rate": 0.00011285034783792416, + "loss": 0.20035157, + "memory(GiB)": 58.24, + "step": 15660, + "train_speed(iter/s)": 0.768503 + }, + { + "epoch": 1.4599627560521415, + "grad_norm": 1.0703125, + "learning_rate": 0.0001126467503305423, + "loss": 0.20264366, + "memory(GiB)": 58.24, + "step": 15680, + "train_speed(iter/s)": 0.768589 + }, + { + "epoch": 1.4618249534450651, + "grad_norm": 1.109375, + "learning_rate": 0.0001124430995338321, + "loss": 0.19577993, + "memory(GiB)": 58.24, + "step": 15700, + "train_speed(iter/s)": 0.76866 + }, + { + "epoch": 1.463687150837989, + "grad_norm": 1.3671875, + "learning_rate": 0.00011223939630591238, + "loss": 0.22569623, + "memory(GiB)": 58.24, + "step": 15720, + "train_speed(iter/s)": 0.768719 + }, + { + "epoch": 1.4655493482309125, + "grad_norm": 1.2578125, + "learning_rate": 0.00011203564150512282, + "loss": 0.20180607, + "memory(GiB)": 58.24, + "step": 15740, + "train_speed(iter/s)": 0.768778 + }, + { + "epoch": 1.4674115456238361, + "grad_norm": 1.2734375, + "learning_rate": 0.00011183183599002052, + "loss": 0.21258569, + "memory(GiB)": 58.24, + "step": 15760, + "train_speed(iter/s)": 0.768849 + }, + { + "epoch": 1.4692737430167597, + "grad_norm": 1.4140625, + "learning_rate": 0.00011162798061937614, + "loss": 0.22123888, + "memory(GiB)": 58.24, + "step": 15780, + "train_speed(iter/s)": 0.768889 + }, + { + "epoch": 1.4711359404096833, + "grad_norm": 1.3984375, + "learning_rate": 0.00011142407625217051, + "loss": 0.20667183, + "memory(GiB)": 58.24, + "step": 15800, + "train_speed(iter/s)": 0.768946 + }, + { + "epoch": 1.4729981378026071, + "grad_norm": 1.6171875, + "learning_rate": 0.0001112201237475909, + "loss": 0.19118414, + "memory(GiB)": 58.24, + "step": 15820, + "train_speed(iter/s)": 0.768992 + }, + { + "epoch": 1.4748603351955307, + "grad_norm": 1.4765625, + "learning_rate": 0.00011101612396502743, + "loss": 0.204529, + "memory(GiB)": 58.24, + "step": 15840, + "train_speed(iter/s)": 0.769069 + }, + { + "epoch": 1.4767225325884543, + "grad_norm": 1.421875, + "learning_rate": 0.00011081207776406936, + "loss": 0.22140031, + "memory(GiB)": 58.24, + "step": 15860, + "train_speed(iter/s)": 0.769122 + }, + { + "epoch": 1.4785847299813781, + "grad_norm": 1.2578125, + "learning_rate": 0.00011060798600450165, + "loss": 0.20754693, + "memory(GiB)": 58.24, + "step": 15880, + "train_speed(iter/s)": 0.769183 + }, + { + "epoch": 1.4804469273743017, + "grad_norm": 1.28125, + "learning_rate": 0.0001104038495463011, + "loss": 0.19853688, + "memory(GiB)": 58.24, + "step": 15900, + "train_speed(iter/s)": 0.76922 + }, + { + "epoch": 1.4823091247672253, + "grad_norm": 1.125, + "learning_rate": 0.00011019966924963301, + "loss": 0.21250277, + "memory(GiB)": 58.24, + "step": 15920, + "train_speed(iter/s)": 0.769279 + }, + { + "epoch": 1.484171322160149, + "grad_norm": 1.5, + "learning_rate": 0.00010999544597484723, + "loss": 0.19000928, + "memory(GiB)": 58.24, + "step": 15940, + "train_speed(iter/s)": 0.769266 + }, + { + "epoch": 1.4860335195530725, + "grad_norm": 1.421875, + "learning_rate": 0.00010979118058247486, + "loss": 0.20164049, + "memory(GiB)": 58.24, + "step": 15960, + "train_speed(iter/s)": 0.769316 + }, + { + "epoch": 1.4878957169459963, + "grad_norm": 1.2578125, + "learning_rate": 0.00010958687393322435, + "loss": 0.18680851, + "memory(GiB)": 58.24, + "step": 15980, + "train_speed(iter/s)": 0.769344 + }, + { + "epoch": 1.48975791433892, + "grad_norm": 1.40625, + "learning_rate": 0.00010938252688797814, + "loss": 0.1877938, + "memory(GiB)": 58.24, + "step": 16000, + "train_speed(iter/s)": 0.769392 + }, + { + "epoch": 1.48975791433892, + "eval_loss": 0.3874351382255554, + "eval_runtime": 76.1306, + "eval_samples_per_second": 182.384, + "eval_steps_per_second": 1.432, + "step": 16000 + }, + { + "epoch": 1.4916201117318435, + "grad_norm": 1.1640625, + "learning_rate": 0.00010917814030778871, + "loss": 0.194278, + "memory(GiB)": 58.24, + "step": 16020, + "train_speed(iter/s)": 0.765509 + }, + { + "epoch": 1.4934823091247673, + "grad_norm": 1.5859375, + "learning_rate": 0.00010897371505387522, + "loss": 0.20364323, + "memory(GiB)": 58.24, + "step": 16040, + "train_speed(iter/s)": 0.765554 + }, + { + "epoch": 1.495344506517691, + "grad_norm": 1.2265625, + "learning_rate": 0.0001087692519876198, + "loss": 0.19808676, + "memory(GiB)": 58.24, + "step": 16060, + "train_speed(iter/s)": 0.765634 + }, + { + "epoch": 1.4972067039106145, + "grad_norm": 1.34375, + "learning_rate": 0.0001085647519705639, + "loss": 0.18924752, + "memory(GiB)": 58.24, + "step": 16080, + "train_speed(iter/s)": 0.765667 + }, + { + "epoch": 1.499068901303538, + "grad_norm": 1.484375, + "learning_rate": 0.00010836021586440461, + "loss": 0.16571351, + "memory(GiB)": 58.24, + "step": 16100, + "train_speed(iter/s)": 0.765745 + }, + { + "epoch": 1.5009310986964617, + "grad_norm": 1.40625, + "learning_rate": 0.00010815564453099116, + "loss": 0.20310044, + "memory(GiB)": 58.24, + "step": 16120, + "train_speed(iter/s)": 0.765771 + }, + { + "epoch": 1.5027932960893855, + "grad_norm": 1.3515625, + "learning_rate": 0.0001079510388323212, + "loss": 0.19717693, + "memory(GiB)": 58.24, + "step": 16140, + "train_speed(iter/s)": 0.765833 + }, + { + "epoch": 1.504655493482309, + "grad_norm": 1.4921875, + "learning_rate": 0.00010774639963053721, + "loss": 0.19102185, + "memory(GiB)": 58.24, + "step": 16160, + "train_speed(iter/s)": 0.765855 + }, + { + "epoch": 1.5065176908752327, + "grad_norm": 1.46875, + "learning_rate": 0.00010754172778792273, + "loss": 0.20653129, + "memory(GiB)": 58.24, + "step": 16180, + "train_speed(iter/s)": 0.765871 + }, + { + "epoch": 1.5083798882681565, + "grad_norm": 1.4921875, + "learning_rate": 0.00010733702416689894, + "loss": 0.17359197, + "memory(GiB)": 58.24, + "step": 16200, + "train_speed(iter/s)": 0.76593 + }, + { + "epoch": 1.51024208566108, + "grad_norm": 1.1484375, + "learning_rate": 0.00010713228963002095, + "loss": 0.20052829, + "memory(GiB)": 58.24, + "step": 16220, + "train_speed(iter/s)": 0.766002 + }, + { + "epoch": 1.5121042830540037, + "grad_norm": 1.2734375, + "learning_rate": 0.00010692752503997405, + "loss": 0.19778749, + "memory(GiB)": 58.24, + "step": 16240, + "train_speed(iter/s)": 0.766039 + }, + { + "epoch": 1.5139664804469275, + "grad_norm": 1.2109375, + "learning_rate": 0.00010672273125957018, + "loss": 0.20196233, + "memory(GiB)": 58.24, + "step": 16260, + "train_speed(iter/s)": 0.766025 + }, + { + "epoch": 1.5158286778398509, + "grad_norm": 1.0546875, + "learning_rate": 0.00010651790915174431, + "loss": 0.18381786, + "memory(GiB)": 58.24, + "step": 16280, + "train_speed(iter/s)": 0.766031 + }, + { + "epoch": 1.5176908752327747, + "grad_norm": 1.0703125, + "learning_rate": 0.0001063130595795508, + "loss": 0.20333216, + "memory(GiB)": 58.24, + "step": 16300, + "train_speed(iter/s)": 0.766055 + }, + { + "epoch": 1.5195530726256983, + "grad_norm": 1.078125, + "learning_rate": 0.00010610818340615967, + "loss": 0.20140872, + "memory(GiB)": 58.24, + "step": 16320, + "train_speed(iter/s)": 0.76611 + }, + { + "epoch": 1.5214152700186219, + "grad_norm": 1.3671875, + "learning_rate": 0.00010590328149485304, + "loss": 0.20071106, + "memory(GiB)": 58.24, + "step": 16340, + "train_speed(iter/s)": 0.76614 + }, + { + "epoch": 1.5232774674115457, + "grad_norm": 1.390625, + "learning_rate": 0.00010569835470902154, + "loss": 0.18391163, + "memory(GiB)": 58.24, + "step": 16360, + "train_speed(iter/s)": 0.766176 + }, + { + "epoch": 1.5251396648044693, + "grad_norm": 1.53125, + "learning_rate": 0.00010549340391216058, + "loss": 0.19652946, + "memory(GiB)": 58.24, + "step": 16380, + "train_speed(iter/s)": 0.76621 + }, + { + "epoch": 1.5270018621973929, + "grad_norm": 1.609375, + "learning_rate": 0.0001052884299678667, + "loss": 0.20681992, + "memory(GiB)": 58.24, + "step": 16400, + "train_speed(iter/s)": 0.766211 + }, + { + "epoch": 1.5288640595903167, + "grad_norm": 0.99609375, + "learning_rate": 0.00010508343373983403, + "loss": 0.17298584, + "memory(GiB)": 58.24, + "step": 16420, + "train_speed(iter/s)": 0.76628 + }, + { + "epoch": 1.5307262569832403, + "grad_norm": 1.3125, + "learning_rate": 0.00010487841609185054, + "loss": 0.20368049, + "memory(GiB)": 58.24, + "step": 16440, + "train_speed(iter/s)": 0.766332 + }, + { + "epoch": 1.5325884543761639, + "grad_norm": 1.6484375, + "learning_rate": 0.00010467337788779456, + "loss": 0.1869107, + "memory(GiB)": 58.24, + "step": 16460, + "train_speed(iter/s)": 0.766398 + }, + { + "epoch": 1.5344506517690877, + "grad_norm": 0.96484375, + "learning_rate": 0.00010446831999163097, + "loss": 0.1824265, + "memory(GiB)": 58.24, + "step": 16480, + "train_speed(iter/s)": 0.766432 + }, + { + "epoch": 1.536312849162011, + "grad_norm": 1.3046875, + "learning_rate": 0.00010426324326740759, + "loss": 0.18272617, + "memory(GiB)": 58.24, + "step": 16500, + "train_speed(iter/s)": 0.766466 + }, + { + "epoch": 1.5381750465549349, + "grad_norm": 1.421875, + "learning_rate": 0.00010405814857925163, + "loss": 0.18423979, + "memory(GiB)": 58.24, + "step": 16520, + "train_speed(iter/s)": 0.766478 + }, + { + "epoch": 1.5400372439478585, + "grad_norm": 1.4296875, + "learning_rate": 0.000103853036791366, + "loss": 0.19179487, + "memory(GiB)": 58.24, + "step": 16540, + "train_speed(iter/s)": 0.766515 + }, + { + "epoch": 1.541899441340782, + "grad_norm": 1.015625, + "learning_rate": 0.00010364790876802564, + "loss": 0.18541435, + "memory(GiB)": 58.24, + "step": 16560, + "train_speed(iter/s)": 0.766542 + }, + { + "epoch": 1.5437616387337059, + "grad_norm": 1.1875, + "learning_rate": 0.00010344276537357385, + "loss": 0.19410855, + "memory(GiB)": 58.24, + "step": 16580, + "train_speed(iter/s)": 0.76658 + }, + { + "epoch": 1.5456238361266295, + "grad_norm": 1.3984375, + "learning_rate": 0.00010323760747241879, + "loss": 0.19085188, + "memory(GiB)": 58.24, + "step": 16600, + "train_speed(iter/s)": 0.766623 + }, + { + "epoch": 1.547486033519553, + "grad_norm": 1.8046875, + "learning_rate": 0.00010303243592902973, + "loss": 0.19986777, + "memory(GiB)": 58.24, + "step": 16620, + "train_speed(iter/s)": 0.766689 + }, + { + "epoch": 1.5493482309124769, + "grad_norm": 1.15625, + "learning_rate": 0.00010282725160793338, + "loss": 0.18039792, + "memory(GiB)": 58.24, + "step": 16640, + "train_speed(iter/s)": 0.76671 + }, + { + "epoch": 1.5512104283054002, + "grad_norm": 1.5703125, + "learning_rate": 0.00010262205537371032, + "loss": 0.18429933, + "memory(GiB)": 58.24, + "step": 16660, + "train_speed(iter/s)": 0.766755 + }, + { + "epoch": 1.553072625698324, + "grad_norm": 1.1953125, + "learning_rate": 0.0001024168480909913, + "loss": 0.18001516, + "memory(GiB)": 58.24, + "step": 16680, + "train_speed(iter/s)": 0.766784 + }, + { + "epoch": 1.5549348230912476, + "grad_norm": 1.2265625, + "learning_rate": 0.00010221163062445367, + "loss": 0.17717788, + "memory(GiB)": 58.24, + "step": 16700, + "train_speed(iter/s)": 0.76683 + }, + { + "epoch": 1.5567970204841712, + "grad_norm": 1.2734375, + "learning_rate": 0.00010200640383881765, + "loss": 0.18292215, + "memory(GiB)": 58.24, + "step": 16720, + "train_speed(iter/s)": 0.766884 + }, + { + "epoch": 1.558659217877095, + "grad_norm": 1.2578125, + "learning_rate": 0.00010180116859884278, + "loss": 0.20042725, + "memory(GiB)": 58.24, + "step": 16740, + "train_speed(iter/s)": 0.766896 + }, + { + "epoch": 1.5605214152700186, + "grad_norm": 1.0625, + "learning_rate": 0.00010159592576932418, + "loss": 0.1987066, + "memory(GiB)": 58.24, + "step": 16760, + "train_speed(iter/s)": 0.766911 + }, + { + "epoch": 1.5623836126629422, + "grad_norm": 1.25, + "learning_rate": 0.0001013906762150889, + "loss": 0.17853866, + "memory(GiB)": 58.24, + "step": 16780, + "train_speed(iter/s)": 0.766976 + }, + { + "epoch": 1.564245810055866, + "grad_norm": 1.1796875, + "learning_rate": 0.00010118542080099247, + "loss": 0.16571226, + "memory(GiB)": 58.24, + "step": 16800, + "train_speed(iter/s)": 0.767027 + }, + { + "epoch": 1.5661080074487894, + "grad_norm": 1.2109375, + "learning_rate": 0.00010098016039191496, + "loss": 0.17807512, + "memory(GiB)": 58.24, + "step": 16820, + "train_speed(iter/s)": 0.767079 + }, + { + "epoch": 1.5679702048417132, + "grad_norm": 1.203125, + "learning_rate": 0.00010077489585275758, + "loss": 0.17454565, + "memory(GiB)": 58.24, + "step": 16840, + "train_speed(iter/s)": 0.767166 + }, + { + "epoch": 1.5698324022346368, + "grad_norm": 1.109375, + "learning_rate": 0.00010056962804843892, + "loss": 0.18413407, + "memory(GiB)": 58.24, + "step": 16860, + "train_speed(iter/s)": 0.767203 + }, + { + "epoch": 1.5716945996275604, + "grad_norm": 1.578125, + "learning_rate": 0.00010036435784389135, + "loss": 0.17773303, + "memory(GiB)": 58.24, + "step": 16880, + "train_speed(iter/s)": 0.767245 + }, + { + "epoch": 1.5735567970204842, + "grad_norm": 1.2734375, + "learning_rate": 0.0001001590861040573, + "loss": 0.19698086, + "memory(GiB)": 58.24, + "step": 16900, + "train_speed(iter/s)": 0.767309 + }, + { + "epoch": 1.5754189944134078, + "grad_norm": 1.1953125, + "learning_rate": 9.99538136938857e-05, + "loss": 0.18003293, + "memory(GiB)": 58.24, + "step": 16920, + "train_speed(iter/s)": 0.767369 + }, + { + "epoch": 1.5772811918063314, + "grad_norm": 1.0859375, + "learning_rate": 9.974854147832826e-05, + "loss": 0.18390901, + "memory(GiB)": 58.24, + "step": 16940, + "train_speed(iter/s)": 0.767439 + }, + { + "epoch": 1.5791433891992552, + "grad_norm": 1.1015625, + "learning_rate": 9.954327032233599e-05, + "loss": 0.17549541, + "memory(GiB)": 58.24, + "step": 16960, + "train_speed(iter/s)": 0.767475 + }, + { + "epoch": 1.5810055865921788, + "grad_norm": 1.1640625, + "learning_rate": 9.933800109085529e-05, + "loss": 0.17035363, + "memory(GiB)": 58.24, + "step": 16980, + "train_speed(iter/s)": 0.767506 + }, + { + "epoch": 1.5828677839851024, + "grad_norm": 1.078125, + "learning_rate": 9.913273464882459e-05, + "loss": 0.16605544, + "memory(GiB)": 58.24, + "step": 17000, + "train_speed(iter/s)": 0.767541 + }, + { + "epoch": 1.5847299813780262, + "grad_norm": 1.40625, + "learning_rate": 9.892747186117047e-05, + "loss": 0.1817858, + "memory(GiB)": 58.24, + "step": 17020, + "train_speed(iter/s)": 0.767592 + }, + { + "epoch": 1.5865921787709496, + "grad_norm": 1.3203125, + "learning_rate": 9.872221359280408e-05, + "loss": 0.17891808, + "memory(GiB)": 58.24, + "step": 17040, + "train_speed(iter/s)": 0.767651 + }, + { + "epoch": 1.5884543761638734, + "grad_norm": 1.296875, + "learning_rate": 9.851696070861766e-05, + "loss": 0.18936658, + "memory(GiB)": 58.24, + "step": 17060, + "train_speed(iter/s)": 0.767704 + }, + { + "epoch": 1.590316573556797, + "grad_norm": 1.296875, + "learning_rate": 9.831171407348064e-05, + "loss": 0.18166518, + "memory(GiB)": 58.24, + "step": 17080, + "train_speed(iter/s)": 0.767746 + }, + { + "epoch": 1.5921787709497206, + "grad_norm": 1.328125, + "learning_rate": 9.810647455223616e-05, + "loss": 0.19375865, + "memory(GiB)": 58.24, + "step": 17100, + "train_speed(iter/s)": 0.767823 + }, + { + "epoch": 1.5940409683426444, + "grad_norm": 1.2578125, + "learning_rate": 9.790124300969743e-05, + "loss": 0.18132524, + "memory(GiB)": 58.24, + "step": 17120, + "train_speed(iter/s)": 0.767869 + }, + { + "epoch": 1.595903165735568, + "grad_norm": 1.21875, + "learning_rate": 9.769602031064394e-05, + "loss": 0.16648118, + "memory(GiB)": 58.24, + "step": 17140, + "train_speed(iter/s)": 0.767933 + }, + { + "epoch": 1.5977653631284916, + "grad_norm": 1.3046875, + "learning_rate": 9.749080731981806e-05, + "loss": 0.17684975, + "memory(GiB)": 58.24, + "step": 17160, + "train_speed(iter/s)": 0.76797 + }, + { + "epoch": 1.5996275605214154, + "grad_norm": 1.34375, + "learning_rate": 9.728560490192112e-05, + "loss": 0.17072924, + "memory(GiB)": 58.24, + "step": 17180, + "train_speed(iter/s)": 0.767983 + }, + { + "epoch": 1.6014897579143388, + "grad_norm": 0.9921875, + "learning_rate": 9.708041392160992e-05, + "loss": 0.17846963, + "memory(GiB)": 58.24, + "step": 17200, + "train_speed(iter/s)": 0.768026 + }, + { + "epoch": 1.6033519553072626, + "grad_norm": 1.4609375, + "learning_rate": 9.687523524349314e-05, + "loss": 0.19516928, + "memory(GiB)": 58.24, + "step": 17220, + "train_speed(iter/s)": 0.768095 + }, + { + "epoch": 1.6052141527001862, + "grad_norm": 1.03125, + "learning_rate": 9.667006973212755e-05, + "loss": 0.16764045, + "memory(GiB)": 58.24, + "step": 17240, + "train_speed(iter/s)": 0.768152 + }, + { + "epoch": 1.6070763500931098, + "grad_norm": 1.09375, + "learning_rate": 9.646491825201447e-05, + "loss": 0.18242341, + "memory(GiB)": 58.24, + "step": 17260, + "train_speed(iter/s)": 0.768197 + }, + { + "epoch": 1.6089385474860336, + "grad_norm": 1.0546875, + "learning_rate": 9.625978166759612e-05, + "loss": 0.17120929, + "memory(GiB)": 58.24, + "step": 17280, + "train_speed(iter/s)": 0.768261 + }, + { + "epoch": 1.6108007448789572, + "grad_norm": 1.1484375, + "learning_rate": 9.605466084325186e-05, + "loss": 0.16303382, + "memory(GiB)": 58.24, + "step": 17300, + "train_speed(iter/s)": 0.768326 + }, + { + "epoch": 1.6126629422718808, + "grad_norm": 1.5234375, + "learning_rate": 9.58495566432948e-05, + "loss": 0.16613436, + "memory(GiB)": 58.24, + "step": 17320, + "train_speed(iter/s)": 0.768393 + }, + { + "epoch": 1.6145251396648046, + "grad_norm": 0.890625, + "learning_rate": 9.564446993196785e-05, + "loss": 0.16981788, + "memory(GiB)": 58.24, + "step": 17340, + "train_speed(iter/s)": 0.768432 + }, + { + "epoch": 1.616387337057728, + "grad_norm": 1.21875, + "learning_rate": 9.543940157344029e-05, + "loss": 0.18024995, + "memory(GiB)": 58.24, + "step": 17360, + "train_speed(iter/s)": 0.768451 + }, + { + "epoch": 1.6182495344506518, + "grad_norm": 1.234375, + "learning_rate": 9.523435243180407e-05, + "loss": 0.18168037, + "memory(GiB)": 58.24, + "step": 17380, + "train_speed(iter/s)": 0.768465 + }, + { + "epoch": 1.6201117318435754, + "grad_norm": 1.0625, + "learning_rate": 9.502932337107018e-05, + "loss": 0.17241347, + "memory(GiB)": 58.24, + "step": 17400, + "train_speed(iter/s)": 0.768438 + }, + { + "epoch": 1.621973929236499, + "grad_norm": 1.3515625, + "learning_rate": 9.482431525516494e-05, + "loss": 0.1620393, + "memory(GiB)": 58.24, + "step": 17420, + "train_speed(iter/s)": 0.768459 + }, + { + "epoch": 1.6238361266294228, + "grad_norm": 1.3125, + "learning_rate": 9.46193289479265e-05, + "loss": 0.18191284, + "memory(GiB)": 58.24, + "step": 17440, + "train_speed(iter/s)": 0.768493 + }, + { + "epoch": 1.6256983240223464, + "grad_norm": 1.3125, + "learning_rate": 9.4414365313101e-05, + "loss": 0.17514572, + "memory(GiB)": 58.24, + "step": 17460, + "train_speed(iter/s)": 0.768567 + }, + { + "epoch": 1.62756052141527, + "grad_norm": 1.3671875, + "learning_rate": 9.420942521433918e-05, + "loss": 0.17852993, + "memory(GiB)": 58.24, + "step": 17480, + "train_speed(iter/s)": 0.768601 + }, + { + "epoch": 1.6294227188081938, + "grad_norm": 1.4140625, + "learning_rate": 9.400450951519253e-05, + "loss": 0.16387472, + "memory(GiB)": 58.24, + "step": 17500, + "train_speed(iter/s)": 0.768606 + }, + { + "epoch": 1.6312849162011172, + "grad_norm": 2.0, + "learning_rate": 9.379961907910967e-05, + "loss": 0.16921017, + "memory(GiB)": 58.24, + "step": 17520, + "train_speed(iter/s)": 0.76869 + }, + { + "epoch": 1.633147113594041, + "grad_norm": 1.1875, + "learning_rate": 9.359475476943292e-05, + "loss": 0.17136528, + "memory(GiB)": 58.24, + "step": 17540, + "train_speed(iter/s)": 0.768707 + }, + { + "epoch": 1.6350093109869648, + "grad_norm": 1.3125, + "learning_rate": 9.338991744939442e-05, + "loss": 0.18011303, + "memory(GiB)": 58.24, + "step": 17560, + "train_speed(iter/s)": 0.768748 + }, + { + "epoch": 1.6368715083798882, + "grad_norm": 1.3046875, + "learning_rate": 9.318510798211253e-05, + "loss": 0.16851079, + "memory(GiB)": 58.24, + "step": 17580, + "train_speed(iter/s)": 0.768783 + }, + { + "epoch": 1.638733705772812, + "grad_norm": 1.3359375, + "learning_rate": 9.298032723058838e-05, + "loss": 0.18129635, + "memory(GiB)": 58.24, + "step": 17600, + "train_speed(iter/s)": 0.768787 + }, + { + "epoch": 1.6405959031657356, + "grad_norm": 1.390625, + "learning_rate": 9.277557605770194e-05, + "loss": 0.15408287, + "memory(GiB)": 58.24, + "step": 17620, + "train_speed(iter/s)": 0.768849 + }, + { + "epoch": 1.6424581005586592, + "grad_norm": 0.99609375, + "learning_rate": 9.257085532620875e-05, + "loss": 0.1833734, + "memory(GiB)": 58.24, + "step": 17640, + "train_speed(iter/s)": 0.768885 + }, + { + "epoch": 1.644320297951583, + "grad_norm": 1.140625, + "learning_rate": 9.236616589873593e-05, + "loss": 0.16197611, + "memory(GiB)": 58.24, + "step": 17660, + "train_speed(iter/s)": 0.768904 + }, + { + "epoch": 1.6461824953445066, + "grad_norm": 1.1484375, + "learning_rate": 9.216150863777865e-05, + "loss": 0.16501439, + "memory(GiB)": 58.24, + "step": 17680, + "train_speed(iter/s)": 0.768969 + }, + { + "epoch": 1.6480446927374302, + "grad_norm": 1.25, + "learning_rate": 9.195688440569675e-05, + "loss": 0.17173784, + "memory(GiB)": 58.24, + "step": 17700, + "train_speed(iter/s)": 0.769007 + }, + { + "epoch": 1.649906890130354, + "grad_norm": 1.3828125, + "learning_rate": 9.175229406471069e-05, + "loss": 0.15976682, + "memory(GiB)": 58.24, + "step": 17720, + "train_speed(iter/s)": 0.769066 + }, + { + "epoch": 1.6517690875232773, + "grad_norm": 1.2578125, + "learning_rate": 9.15477384768982e-05, + "loss": 0.16614034, + "memory(GiB)": 58.24, + "step": 17740, + "train_speed(iter/s)": 0.769113 + }, + { + "epoch": 1.6536312849162011, + "grad_norm": 1.1171875, + "learning_rate": 9.134321850419062e-05, + "loss": 0.19003222, + "memory(GiB)": 58.24, + "step": 17760, + "train_speed(iter/s)": 0.769151 + }, + { + "epoch": 1.6554934823091247, + "grad_norm": 1.0078125, + "learning_rate": 9.11387350083691e-05, + "loss": 0.17415506, + "memory(GiB)": 58.24, + "step": 17780, + "train_speed(iter/s)": 0.769173 + }, + { + "epoch": 1.6573556797020483, + "grad_norm": 1.203125, + "learning_rate": 9.093428885106127e-05, + "loss": 0.16781898, + "memory(GiB)": 58.24, + "step": 17800, + "train_speed(iter/s)": 0.769224 + }, + { + "epoch": 1.6592178770949721, + "grad_norm": 1.15625, + "learning_rate": 9.072988089373724e-05, + "loss": 0.16520228, + "memory(GiB)": 58.24, + "step": 17820, + "train_speed(iter/s)": 0.769277 + }, + { + "epoch": 1.6610800744878957, + "grad_norm": 1.578125, + "learning_rate": 9.052551199770623e-05, + "loss": 0.17041247, + "memory(GiB)": 58.24, + "step": 17840, + "train_speed(iter/s)": 0.769331 + }, + { + "epoch": 1.6629422718808193, + "grad_norm": 1.09375, + "learning_rate": 9.032118302411293e-05, + "loss": 0.18079939, + "memory(GiB)": 58.24, + "step": 17860, + "train_speed(iter/s)": 0.769372 + }, + { + "epoch": 1.6648044692737431, + "grad_norm": 1.015625, + "learning_rate": 9.011689483393375e-05, + "loss": 0.15872812, + "memory(GiB)": 58.24, + "step": 17880, + "train_speed(iter/s)": 0.76943 + }, + { + "epoch": 1.6666666666666665, + "grad_norm": 1.515625, + "learning_rate": 8.991264828797319e-05, + "loss": 0.16760997, + "memory(GiB)": 58.24, + "step": 17900, + "train_speed(iter/s)": 0.769459 + }, + { + "epoch": 1.6685288640595903, + "grad_norm": 1.125, + "learning_rate": 8.970844424686041e-05, + "loss": 0.16810267, + "memory(GiB)": 58.24, + "step": 17920, + "train_speed(iter/s)": 0.769493 + }, + { + "epoch": 1.670391061452514, + "grad_norm": 1.25, + "learning_rate": 8.950428357104535e-05, + "loss": 0.16381236, + "memory(GiB)": 58.24, + "step": 17940, + "train_speed(iter/s)": 0.769526 + }, + { + "epoch": 1.6722532588454375, + "grad_norm": 0.8828125, + "learning_rate": 8.930016712079531e-05, + "loss": 0.15553318, + "memory(GiB)": 58.24, + "step": 17960, + "train_speed(iter/s)": 0.76956 + }, + { + "epoch": 1.6741154562383613, + "grad_norm": 0.96875, + "learning_rate": 8.909609575619121e-05, + "loss": 0.17401633, + "memory(GiB)": 58.24, + "step": 17980, + "train_speed(iter/s)": 0.76963 + }, + { + "epoch": 1.675977653631285, + "grad_norm": 1.5703125, + "learning_rate": 8.889207033712392e-05, + "loss": 0.16966208, + "memory(GiB)": 58.24, + "step": 18000, + "train_speed(iter/s)": 0.76966 + }, + { + "epoch": 1.675977653631285, + "eval_loss": 0.3787587583065033, + "eval_runtime": 76.3917, + "eval_samples_per_second": 181.761, + "eval_steps_per_second": 1.427, + "step": 18000 + }, + { + "epoch": 1.6778398510242085, + "grad_norm": 1.140625, + "learning_rate": 8.868809172329084e-05, + "loss": 0.16671758, + "memory(GiB)": 58.24, + "step": 18020, + "train_speed(iter/s)": 0.766139 + }, + { + "epoch": 1.6797020484171323, + "grad_norm": 1.15625, + "learning_rate": 8.848416077419208e-05, + "loss": 0.1657429, + "memory(GiB)": 58.24, + "step": 18040, + "train_speed(iter/s)": 0.766174 + }, + { + "epoch": 1.6815642458100557, + "grad_norm": 1.359375, + "learning_rate": 8.828027834912686e-05, + "loss": 0.16353419, + "memory(GiB)": 58.24, + "step": 18060, + "train_speed(iter/s)": 0.766179 + }, + { + "epoch": 1.6834264432029795, + "grad_norm": 1.28125, + "learning_rate": 8.807644530719002e-05, + "loss": 0.15319026, + "memory(GiB)": 58.24, + "step": 18080, + "train_speed(iter/s)": 0.766193 + }, + { + "epoch": 1.6852886405959033, + "grad_norm": 1.171875, + "learning_rate": 8.787266250726824e-05, + "loss": 0.15553379, + "memory(GiB)": 58.24, + "step": 18100, + "train_speed(iter/s)": 0.76623 + }, + { + "epoch": 1.6871508379888267, + "grad_norm": 1.40625, + "learning_rate": 8.766893080803659e-05, + "loss": 0.16313032, + "memory(GiB)": 58.24, + "step": 18120, + "train_speed(iter/s)": 0.766265 + }, + { + "epoch": 1.6890130353817505, + "grad_norm": 1.140625, + "learning_rate": 8.746525106795476e-05, + "loss": 0.16240375, + "memory(GiB)": 58.24, + "step": 18140, + "train_speed(iter/s)": 0.766296 + }, + { + "epoch": 1.690875232774674, + "grad_norm": 1.234375, + "learning_rate": 8.726162414526343e-05, + "loss": 0.14353172, + "memory(GiB)": 58.24, + "step": 18160, + "train_speed(iter/s)": 0.766374 + }, + { + "epoch": 1.6927374301675977, + "grad_norm": 1.203125, + "learning_rate": 8.705805089798089e-05, + "loss": 0.15949094, + "memory(GiB)": 58.24, + "step": 18180, + "train_speed(iter/s)": 0.766413 + }, + { + "epoch": 1.6945996275605215, + "grad_norm": 1.1484375, + "learning_rate": 8.685453218389914e-05, + "loss": 0.15888497, + "memory(GiB)": 58.24, + "step": 18200, + "train_speed(iter/s)": 0.76646 + }, + { + "epoch": 1.696461824953445, + "grad_norm": 1.265625, + "learning_rate": 8.665106886058037e-05, + "loss": 0.16196404, + "memory(GiB)": 58.24, + "step": 18220, + "train_speed(iter/s)": 0.766503 + }, + { + "epoch": 1.6983240223463687, + "grad_norm": 1.1484375, + "learning_rate": 8.644766178535351e-05, + "loss": 0.15065837, + "memory(GiB)": 58.24, + "step": 18240, + "train_speed(iter/s)": 0.766528 + }, + { + "epoch": 1.7001862197392925, + "grad_norm": 1.25, + "learning_rate": 8.624431181531029e-05, + "loss": 0.15646677, + "memory(GiB)": 58.24, + "step": 18260, + "train_speed(iter/s)": 0.766571 + }, + { + "epoch": 1.7020484171322159, + "grad_norm": 1.2578125, + "learning_rate": 8.604101980730202e-05, + "loss": 0.17705973, + "memory(GiB)": 58.24, + "step": 18280, + "train_speed(iter/s)": 0.766602 + }, + { + "epoch": 1.7039106145251397, + "grad_norm": 0.96875, + "learning_rate": 8.583778661793565e-05, + "loss": 0.1656255, + "memory(GiB)": 58.24, + "step": 18300, + "train_speed(iter/s)": 0.76666 + }, + { + "epoch": 1.7057728119180633, + "grad_norm": 1.4765625, + "learning_rate": 8.563461310357023e-05, + "loss": 0.16021589, + "memory(GiB)": 58.24, + "step": 18320, + "train_speed(iter/s)": 0.766706 + }, + { + "epoch": 1.7076350093109869, + "grad_norm": 1.4921875, + "learning_rate": 8.543150012031357e-05, + "loss": 0.15275285, + "memory(GiB)": 58.24, + "step": 18340, + "train_speed(iter/s)": 0.766732 + }, + { + "epoch": 1.7094972067039107, + "grad_norm": 1.1484375, + "learning_rate": 8.522844852401824e-05, + "loss": 0.16219683, + "memory(GiB)": 58.24, + "step": 18360, + "train_speed(iter/s)": 0.766794 + }, + { + "epoch": 1.7113594040968343, + "grad_norm": 1.09375, + "learning_rate": 8.502545917027817e-05, + "loss": 0.1585273, + "memory(GiB)": 58.24, + "step": 18380, + "train_speed(iter/s)": 0.7668 + }, + { + "epoch": 1.7132216014897579, + "grad_norm": 0.8359375, + "learning_rate": 8.482253291442508e-05, + "loss": 0.15480431, + "memory(GiB)": 58.24, + "step": 18400, + "train_speed(iter/s)": 0.766861 + }, + { + "epoch": 1.7150837988826817, + "grad_norm": 1.125, + "learning_rate": 8.461967061152475e-05, + "loss": 0.16871202, + "memory(GiB)": 58.24, + "step": 18420, + "train_speed(iter/s)": 0.766888 + }, + { + "epoch": 1.716945996275605, + "grad_norm": 1.203125, + "learning_rate": 8.441687311637358e-05, + "loss": 0.14869682, + "memory(GiB)": 58.24, + "step": 18440, + "train_speed(iter/s)": 0.766939 + }, + { + "epoch": 1.7188081936685289, + "grad_norm": 1.0546875, + "learning_rate": 8.421414128349478e-05, + "loss": 0.15027819, + "memory(GiB)": 58.24, + "step": 18460, + "train_speed(iter/s)": 0.76699 + }, + { + "epoch": 1.7206703910614525, + "grad_norm": 1.1171875, + "learning_rate": 8.401147596713492e-05, + "loss": 0.1574163, + "memory(GiB)": 58.24, + "step": 18480, + "train_speed(iter/s)": 0.767002 + }, + { + "epoch": 1.722532588454376, + "grad_norm": 1.0703125, + "learning_rate": 8.380887802126036e-05, + "loss": 0.1544909, + "memory(GiB)": 58.24, + "step": 18500, + "train_speed(iter/s)": 0.767039 + }, + { + "epoch": 1.7243947858472999, + "grad_norm": 1.1484375, + "learning_rate": 8.360634829955349e-05, + "loss": 0.14190108, + "memory(GiB)": 58.24, + "step": 18520, + "train_speed(iter/s)": 0.767069 + }, + { + "epoch": 1.7262569832402235, + "grad_norm": 1.1328125, + "learning_rate": 8.340388765540923e-05, + "loss": 0.15983729, + "memory(GiB)": 58.24, + "step": 18540, + "train_speed(iter/s)": 0.76707 + }, + { + "epoch": 1.728119180633147, + "grad_norm": 1.3046875, + "learning_rate": 8.320149694193147e-05, + "loss": 0.16751406, + "memory(GiB)": 58.24, + "step": 18560, + "train_speed(iter/s)": 0.767086 + }, + { + "epoch": 1.7299813780260709, + "grad_norm": 1.3125, + "learning_rate": 8.299917701192943e-05, + "loss": 0.16194601, + "memory(GiB)": 58.24, + "step": 18580, + "train_speed(iter/s)": 0.767106 + }, + { + "epoch": 1.7318435754189943, + "grad_norm": 1.1796875, + "learning_rate": 8.279692871791411e-05, + "loss": 0.16382195, + "memory(GiB)": 58.24, + "step": 18600, + "train_speed(iter/s)": 0.767163 + }, + { + "epoch": 1.733705772811918, + "grad_norm": 1.40625, + "learning_rate": 8.259475291209454e-05, + "loss": 0.15488023, + "memory(GiB)": 58.24, + "step": 18620, + "train_speed(iter/s)": 0.767205 + }, + { + "epoch": 1.7355679702048417, + "grad_norm": 1.25, + "learning_rate": 8.239265044637438e-05, + "loss": 0.1414247, + "memory(GiB)": 58.24, + "step": 18640, + "train_speed(iter/s)": 0.767229 + }, + { + "epoch": 1.7374301675977653, + "grad_norm": 1.2890625, + "learning_rate": 8.219062217234832e-05, + "loss": 0.15719596, + "memory(GiB)": 58.24, + "step": 18660, + "train_speed(iter/s)": 0.76727 + }, + { + "epoch": 1.739292364990689, + "grad_norm": 1.28125, + "learning_rate": 8.198866894129832e-05, + "loss": 0.15911665, + "memory(GiB)": 58.24, + "step": 18680, + "train_speed(iter/s)": 0.767339 + }, + { + "epoch": 1.7411545623836127, + "grad_norm": 1.0, + "learning_rate": 8.178679160419018e-05, + "loss": 0.14015193, + "memory(GiB)": 58.24, + "step": 18700, + "train_speed(iter/s)": 0.767386 + }, + { + "epoch": 1.7430167597765363, + "grad_norm": 1.2890625, + "learning_rate": 8.158499101166997e-05, + "loss": 0.16646228, + "memory(GiB)": 58.24, + "step": 18720, + "train_speed(iter/s)": 0.767387 + }, + { + "epoch": 1.74487895716946, + "grad_norm": 1.015625, + "learning_rate": 8.138326801406025e-05, + "loss": 0.13748178, + "memory(GiB)": 58.24, + "step": 18740, + "train_speed(iter/s)": 0.767449 + }, + { + "epoch": 1.7467411545623837, + "grad_norm": 1.2109375, + "learning_rate": 8.118162346135674e-05, + "loss": 0.16049628, + "memory(GiB)": 58.24, + "step": 18760, + "train_speed(iter/s)": 0.767496 + }, + { + "epoch": 1.7486033519553073, + "grad_norm": 0.92578125, + "learning_rate": 8.098005820322455e-05, + "loss": 0.14880733, + "memory(GiB)": 58.24, + "step": 18780, + "train_speed(iter/s)": 0.767549 + }, + { + "epoch": 1.750465549348231, + "grad_norm": 1.078125, + "learning_rate": 8.077857308899465e-05, + "loss": 0.16304884, + "memory(GiB)": 58.24, + "step": 18800, + "train_speed(iter/s)": 0.767576 + }, + { + "epoch": 1.7523277467411544, + "grad_norm": 1.1640625, + "learning_rate": 8.057716896766044e-05, + "loss": 0.14319315, + "memory(GiB)": 58.24, + "step": 18820, + "train_speed(iter/s)": 0.767627 + }, + { + "epoch": 1.7541899441340782, + "grad_norm": 1.34375, + "learning_rate": 8.037584668787391e-05, + "loss": 0.1478158, + "memory(GiB)": 58.24, + "step": 18840, + "train_speed(iter/s)": 0.767631 + }, + { + "epoch": 1.7560521415270018, + "grad_norm": 1.1484375, + "learning_rate": 8.017460709794216e-05, + "loss": 0.1471365, + "memory(GiB)": 58.24, + "step": 18860, + "train_speed(iter/s)": 0.767666 + }, + { + "epoch": 1.7579143389199254, + "grad_norm": 1.140625, + "learning_rate": 7.997345104582405e-05, + "loss": 0.13758318, + "memory(GiB)": 58.24, + "step": 18880, + "train_speed(iter/s)": 0.767694 + }, + { + "epoch": 1.7597765363128492, + "grad_norm": 1.265625, + "learning_rate": 7.97723793791263e-05, + "loss": 0.14244239, + "memory(GiB)": 58.24, + "step": 18900, + "train_speed(iter/s)": 0.767758 + }, + { + "epoch": 1.7616387337057728, + "grad_norm": 0.984375, + "learning_rate": 7.957139294510003e-05, + "loss": 0.15055372, + "memory(GiB)": 58.24, + "step": 18920, + "train_speed(iter/s)": 0.767801 + }, + { + "epoch": 1.7635009310986964, + "grad_norm": 1.109375, + "learning_rate": 7.937049259063734e-05, + "loss": 0.15460105, + "memory(GiB)": 58.24, + "step": 18940, + "train_speed(iter/s)": 0.767834 + }, + { + "epoch": 1.7653631284916202, + "grad_norm": 1.1484375, + "learning_rate": 7.916967916226747e-05, + "loss": 0.13544981, + "memory(GiB)": 58.24, + "step": 18960, + "train_speed(iter/s)": 0.767892 + }, + { + "epoch": 1.7672253258845436, + "grad_norm": 1.40625, + "learning_rate": 7.896895350615357e-05, + "loss": 0.15087662, + "memory(GiB)": 58.24, + "step": 18980, + "train_speed(iter/s)": 0.76795 + }, + { + "epoch": 1.7690875232774674, + "grad_norm": 0.8671875, + "learning_rate": 7.876831646808874e-05, + "loss": 0.13152529, + "memory(GiB)": 58.24, + "step": 19000, + "train_speed(iter/s)": 0.767981 + }, + { + "epoch": 1.770949720670391, + "grad_norm": 1.3515625, + "learning_rate": 7.856776889349279e-05, + "loss": 0.1433506, + "memory(GiB)": 58.24, + "step": 19020, + "train_speed(iter/s)": 0.768017 + }, + { + "epoch": 1.7728119180633146, + "grad_norm": 1.3203125, + "learning_rate": 7.836731162740857e-05, + "loss": 0.16239972, + "memory(GiB)": 58.24, + "step": 19040, + "train_speed(iter/s)": 0.768055 + }, + { + "epoch": 1.7746741154562384, + "grad_norm": 1.0546875, + "learning_rate": 7.816694551449838e-05, + "loss": 0.15868897, + "memory(GiB)": 58.24, + "step": 19060, + "train_speed(iter/s)": 0.768099 + }, + { + "epoch": 1.776536312849162, + "grad_norm": 1.15625, + "learning_rate": 7.796667139904036e-05, + "loss": 0.13739709, + "memory(GiB)": 58.24, + "step": 19080, + "train_speed(iter/s)": 0.768138 + }, + { + "epoch": 1.7783985102420856, + "grad_norm": 1.2734375, + "learning_rate": 7.77664901249251e-05, + "loss": 0.14739815, + "memory(GiB)": 58.24, + "step": 19100, + "train_speed(iter/s)": 0.768153 + }, + { + "epoch": 1.7802607076350094, + "grad_norm": 1.0859375, + "learning_rate": 7.75664025356519e-05, + "loss": 0.15729291, + "memory(GiB)": 58.24, + "step": 19120, + "train_speed(iter/s)": 0.768177 + }, + { + "epoch": 1.7821229050279328, + "grad_norm": 1.3359375, + "learning_rate": 7.736640947432544e-05, + "loss": 0.14836711, + "memory(GiB)": 58.24, + "step": 19140, + "train_speed(iter/s)": 0.768201 + }, + { + "epoch": 1.7839851024208566, + "grad_norm": 1.0859375, + "learning_rate": 7.716651178365193e-05, + "loss": 0.1370422, + "memory(GiB)": 58.24, + "step": 19160, + "train_speed(iter/s)": 0.768264 + }, + { + "epoch": 1.7858472998137802, + "grad_norm": 1.1640625, + "learning_rate": 7.696671030593575e-05, + "loss": 0.13977002, + "memory(GiB)": 58.24, + "step": 19180, + "train_speed(iter/s)": 0.768294 + }, + { + "epoch": 1.7877094972067038, + "grad_norm": 0.953125, + "learning_rate": 7.676700588307601e-05, + "loss": 0.13385307, + "memory(GiB)": 58.24, + "step": 19200, + "train_speed(iter/s)": 0.768336 + }, + { + "epoch": 1.7895716945996276, + "grad_norm": 1.421875, + "learning_rate": 7.656739935656271e-05, + "loss": 0.15971838, + "memory(GiB)": 58.24, + "step": 19220, + "train_speed(iter/s)": 0.768358 + }, + { + "epoch": 1.7914338919925512, + "grad_norm": 0.9609375, + "learning_rate": 7.636789156747338e-05, + "loss": 0.14956834, + "memory(GiB)": 58.24, + "step": 19240, + "train_speed(iter/s)": 0.76839 + }, + { + "epoch": 1.7932960893854748, + "grad_norm": 1.34375, + "learning_rate": 7.616848335646956e-05, + "loss": 0.14564731, + "memory(GiB)": 58.24, + "step": 19260, + "train_speed(iter/s)": 0.76842 + }, + { + "epoch": 1.7951582867783986, + "grad_norm": 1.3125, + "learning_rate": 7.596917556379305e-05, + "loss": 0.14346001, + "memory(GiB)": 58.24, + "step": 19280, + "train_speed(iter/s)": 0.768463 + }, + { + "epoch": 1.7970204841713222, + "grad_norm": 1.015625, + "learning_rate": 7.576996902926278e-05, + "loss": 0.14654245, + "memory(GiB)": 58.24, + "step": 19300, + "train_speed(iter/s)": 0.768515 + }, + { + "epoch": 1.7988826815642458, + "grad_norm": 1.1953125, + "learning_rate": 7.557086459227078e-05, + "loss": 0.15021694, + "memory(GiB)": 58.24, + "step": 19320, + "train_speed(iter/s)": 0.768549 + }, + { + "epoch": 1.8007448789571696, + "grad_norm": 1.1171875, + "learning_rate": 7.537186309177892e-05, + "loss": 0.15023569, + "memory(GiB)": 58.24, + "step": 19340, + "train_speed(iter/s)": 0.768581 + }, + { + "epoch": 1.802607076350093, + "grad_norm": 1.2890625, + "learning_rate": 7.517296536631544e-05, + "loss": 0.15608531, + "memory(GiB)": 58.24, + "step": 19360, + "train_speed(iter/s)": 0.768616 + }, + { + "epoch": 1.8044692737430168, + "grad_norm": 1.0546875, + "learning_rate": 7.497417225397118e-05, + "loss": 0.15757583, + "memory(GiB)": 58.24, + "step": 19380, + "train_speed(iter/s)": 0.768653 + }, + { + "epoch": 1.8063314711359404, + "grad_norm": 1.046875, + "learning_rate": 7.477548459239623e-05, + "loss": 0.14360132, + "memory(GiB)": 58.24, + "step": 19400, + "train_speed(iter/s)": 0.768689 + }, + { + "epoch": 1.808193668528864, + "grad_norm": 1.3671875, + "learning_rate": 7.457690321879632e-05, + "loss": 0.15476027, + "memory(GiB)": 58.24, + "step": 19420, + "train_speed(iter/s)": 0.768724 + }, + { + "epoch": 1.8100558659217878, + "grad_norm": 1.21875, + "learning_rate": 7.437842896992932e-05, + "loss": 0.13898827, + "memory(GiB)": 58.24, + "step": 19440, + "train_speed(iter/s)": 0.768768 + }, + { + "epoch": 1.8119180633147114, + "grad_norm": 1.109375, + "learning_rate": 7.418006268210178e-05, + "loss": 0.14609106, + "memory(GiB)": 58.24, + "step": 19460, + "train_speed(iter/s)": 0.768823 + }, + { + "epoch": 1.813780260707635, + "grad_norm": 1.359375, + "learning_rate": 7.398180519116526e-05, + "loss": 0.14926387, + "memory(GiB)": 58.24, + "step": 19480, + "train_speed(iter/s)": 0.768889 + }, + { + "epoch": 1.8156424581005588, + "grad_norm": 1.2578125, + "learning_rate": 7.378365733251284e-05, + "loss": 0.14744906, + "memory(GiB)": 58.24, + "step": 19500, + "train_speed(iter/s)": 0.768922 + }, + { + "epoch": 1.8175046554934822, + "grad_norm": 1.2890625, + "learning_rate": 7.35856199410758e-05, + "loss": 0.15171263, + "memory(GiB)": 58.24, + "step": 19520, + "train_speed(iter/s)": 0.768968 + }, + { + "epoch": 1.819366852886406, + "grad_norm": 1.2109375, + "learning_rate": 7.33876938513198e-05, + "loss": 0.14976683, + "memory(GiB)": 58.24, + "step": 19540, + "train_speed(iter/s)": 0.769005 + }, + { + "epoch": 1.8212290502793296, + "grad_norm": 1.4609375, + "learning_rate": 7.31898798972416e-05, + "loss": 0.13972685, + "memory(GiB)": 58.24, + "step": 19560, + "train_speed(iter/s)": 0.769011 + }, + { + "epoch": 1.8230912476722532, + "grad_norm": 1.2421875, + "learning_rate": 7.29921789123654e-05, + "loss": 0.14240303, + "memory(GiB)": 58.24, + "step": 19580, + "train_speed(iter/s)": 0.769071 + }, + { + "epoch": 1.824953445065177, + "grad_norm": 1.1328125, + "learning_rate": 7.27945917297394e-05, + "loss": 0.15212356, + "memory(GiB)": 58.24, + "step": 19600, + "train_speed(iter/s)": 0.769081 + }, + { + "epoch": 1.8268156424581006, + "grad_norm": 1.4609375, + "learning_rate": 7.25971191819323e-05, + "loss": 0.13351725, + "memory(GiB)": 58.24, + "step": 19620, + "train_speed(iter/s)": 0.769122 + }, + { + "epoch": 1.8286778398510242, + "grad_norm": 1.203125, + "learning_rate": 7.239976210102977e-05, + "loss": 0.14566996, + "memory(GiB)": 58.24, + "step": 19640, + "train_speed(iter/s)": 0.769148 + }, + { + "epoch": 1.830540037243948, + "grad_norm": 0.93359375, + "learning_rate": 7.220252131863084e-05, + "loss": 0.1393986, + "memory(GiB)": 58.24, + "step": 19660, + "train_speed(iter/s)": 0.769214 + }, + { + "epoch": 1.8324022346368714, + "grad_norm": 0.96484375, + "learning_rate": 7.200539766584467e-05, + "loss": 0.14144776, + "memory(GiB)": 58.24, + "step": 19680, + "train_speed(iter/s)": 0.76923 + }, + { + "epoch": 1.8342644320297952, + "grad_norm": 0.87109375, + "learning_rate": 7.180839197328674e-05, + "loss": 0.13021162, + "memory(GiB)": 58.24, + "step": 19700, + "train_speed(iter/s)": 0.769244 + }, + { + "epoch": 1.8361266294227188, + "grad_norm": 1.046875, + "learning_rate": 7.161150507107547e-05, + "loss": 0.14698963, + "memory(GiB)": 58.24, + "step": 19720, + "train_speed(iter/s)": 0.769265 + }, + { + "epoch": 1.8379888268156424, + "grad_norm": 1.140625, + "learning_rate": 7.141473778882883e-05, + "loss": 0.1384059, + "memory(GiB)": 58.24, + "step": 19740, + "train_speed(iter/s)": 0.769284 + }, + { + "epoch": 1.8398510242085662, + "grad_norm": 1.3203125, + "learning_rate": 7.121809095566072e-05, + "loss": 0.13622018, + "memory(GiB)": 58.24, + "step": 19760, + "train_speed(iter/s)": 0.769313 + }, + { + "epoch": 1.8417132216014898, + "grad_norm": 0.8828125, + "learning_rate": 7.102156540017748e-05, + "loss": 0.13797989, + "memory(GiB)": 58.24, + "step": 19780, + "train_speed(iter/s)": 0.769333 + }, + { + "epoch": 1.8435754189944134, + "grad_norm": 1.1015625, + "learning_rate": 7.082516195047443e-05, + "loss": 0.15423731, + "memory(GiB)": 58.24, + "step": 19800, + "train_speed(iter/s)": 0.769378 + }, + { + "epoch": 1.8454376163873372, + "grad_norm": 1.390625, + "learning_rate": 7.062888143413235e-05, + "loss": 0.14557085, + "memory(GiB)": 58.24, + "step": 19820, + "train_speed(iter/s)": 0.769413 + }, + { + "epoch": 1.8472998137802608, + "grad_norm": 0.875, + "learning_rate": 7.043272467821412e-05, + "loss": 0.11753706, + "memory(GiB)": 58.24, + "step": 19840, + "train_speed(iter/s)": 0.769451 + }, + { + "epoch": 1.8491620111731844, + "grad_norm": 1.28125, + "learning_rate": 7.023669250926103e-05, + "loss": 0.13218291, + "memory(GiB)": 58.24, + "step": 19860, + "train_speed(iter/s)": 0.769492 + }, + { + "epoch": 1.8510242085661082, + "grad_norm": 1.1484375, + "learning_rate": 7.00407857532894e-05, + "loss": 0.13124348, + "memory(GiB)": 58.24, + "step": 19880, + "train_speed(iter/s)": 0.769509 + }, + { + "epoch": 1.8528864059590315, + "grad_norm": 1.1015625, + "learning_rate": 6.98450052357872e-05, + "loss": 0.1268985, + "memory(GiB)": 58.24, + "step": 19900, + "train_speed(iter/s)": 0.769551 + }, + { + "epoch": 1.8547486033519553, + "grad_norm": 0.9765625, + "learning_rate": 6.964935178171037e-05, + "loss": 0.13166354, + "memory(GiB)": 58.24, + "step": 19920, + "train_speed(iter/s)": 0.769587 + }, + { + "epoch": 1.856610800744879, + "grad_norm": 1.046875, + "learning_rate": 6.945382621547951e-05, + "loss": 0.13858917, + "memory(GiB)": 58.24, + "step": 19940, + "train_speed(iter/s)": 0.769637 + }, + { + "epoch": 1.8584729981378025, + "grad_norm": 1.2265625, + "learning_rate": 6.925842936097629e-05, + "loss": 0.14260222, + "memory(GiB)": 58.24, + "step": 19960, + "train_speed(iter/s)": 0.769684 + }, + { + "epoch": 1.8603351955307263, + "grad_norm": 1.0625, + "learning_rate": 6.906316204154002e-05, + "loss": 0.14681377, + "memory(GiB)": 58.24, + "step": 19980, + "train_speed(iter/s)": 0.769741 + }, + { + "epoch": 1.86219739292365, + "grad_norm": 1.3671875, + "learning_rate": 6.886802507996432e-05, + "loss": 0.1321399, + "memory(GiB)": 58.24, + "step": 20000, + "train_speed(iter/s)": 0.769748 + }, + { + "epoch": 1.86219739292365, + "eval_loss": 0.3764718770980835, + "eval_runtime": 77.0126, + "eval_samples_per_second": 180.295, + "eval_steps_per_second": 1.415, + "step": 20000 + }, + { + "epoch": 1.8640595903165735, + "grad_norm": 1.109375, + "learning_rate": 6.867301929849337e-05, + "loss": 0.13974258, + "memory(GiB)": 58.24, + "step": 20020, + "train_speed(iter/s)": 0.766626 + }, + { + "epoch": 1.8659217877094973, + "grad_norm": 1.1015625, + "learning_rate": 6.847814551881861e-05, + "loss": 0.13684764, + "memory(GiB)": 58.24, + "step": 20040, + "train_speed(iter/s)": 0.766677 + }, + { + "epoch": 1.8677839851024207, + "grad_norm": 1.390625, + "learning_rate": 6.828340456207542e-05, + "loss": 0.13937807, + "memory(GiB)": 58.24, + "step": 20060, + "train_speed(iter/s)": 0.766727 + }, + { + "epoch": 1.8696461824953445, + "grad_norm": 1.390625, + "learning_rate": 6.808879724883935e-05, + "loss": 0.14577636, + "memory(GiB)": 58.24, + "step": 20080, + "train_speed(iter/s)": 0.766742 + }, + { + "epoch": 1.8715083798882681, + "grad_norm": 1.109375, + "learning_rate": 6.789432439912289e-05, + "loss": 0.12374585, + "memory(GiB)": 58.24, + "step": 20100, + "train_speed(iter/s)": 0.766781 + }, + { + "epoch": 1.8733705772811917, + "grad_norm": 1.046875, + "learning_rate": 6.769998683237193e-05, + "loss": 0.12786458, + "memory(GiB)": 58.24, + "step": 20120, + "train_speed(iter/s)": 0.766806 + }, + { + "epoch": 1.8752327746741155, + "grad_norm": 1.1875, + "learning_rate": 6.750578536746224e-05, + "loss": 0.12823834, + "memory(GiB)": 58.24, + "step": 20140, + "train_speed(iter/s)": 0.766836 + }, + { + "epoch": 1.8770949720670391, + "grad_norm": 1.6171875, + "learning_rate": 6.73117208226963e-05, + "loss": 0.14647946, + "memory(GiB)": 58.24, + "step": 20160, + "train_speed(iter/s)": 0.766845 + }, + { + "epoch": 1.8789571694599627, + "grad_norm": 0.9921875, + "learning_rate": 6.711779401579947e-05, + "loss": 0.12733679, + "memory(GiB)": 58.24, + "step": 20180, + "train_speed(iter/s)": 0.766908 + }, + { + "epoch": 1.8808193668528865, + "grad_norm": 1.1484375, + "learning_rate": 6.692400576391675e-05, + "loss": 0.1478996, + "memory(GiB)": 58.24, + "step": 20200, + "train_speed(iter/s)": 0.76694 + }, + { + "epoch": 1.88268156424581, + "grad_norm": 1.3203125, + "learning_rate": 6.673035688360944e-05, + "loss": 0.13632859, + "memory(GiB)": 58.24, + "step": 20220, + "train_speed(iter/s)": 0.767002 + }, + { + "epoch": 1.8845437616387337, + "grad_norm": 1.4921875, + "learning_rate": 6.653684819085142e-05, + "loss": 0.13835206, + "memory(GiB)": 58.24, + "step": 20240, + "train_speed(iter/s)": 0.767049 + }, + { + "epoch": 1.8864059590316573, + "grad_norm": 2.453125, + "learning_rate": 6.634348050102597e-05, + "loss": 0.15622594, + "memory(GiB)": 58.24, + "step": 20260, + "train_speed(iter/s)": 0.76704 + }, + { + "epoch": 1.888268156424581, + "grad_norm": 1.0, + "learning_rate": 6.615025462892218e-05, + "loss": 0.12626829, + "memory(GiB)": 58.24, + "step": 20280, + "train_speed(iter/s)": 0.767074 + }, + { + "epoch": 1.8901303538175047, + "grad_norm": 1.1796875, + "learning_rate": 6.59571713887315e-05, + "loss": 0.13793883, + "memory(GiB)": 58.24, + "step": 20300, + "train_speed(iter/s)": 0.767106 + }, + { + "epoch": 1.8919925512104283, + "grad_norm": 1.390625, + "learning_rate": 6.576423159404456e-05, + "loss": 0.12859577, + "memory(GiB)": 58.24, + "step": 20320, + "train_speed(iter/s)": 0.767136 + }, + { + "epoch": 1.893854748603352, + "grad_norm": 1.4296875, + "learning_rate": 6.557143605784742e-05, + "loss": 0.13382497, + "memory(GiB)": 58.24, + "step": 20340, + "train_speed(iter/s)": 0.767179 + }, + { + "epoch": 1.8957169459962757, + "grad_norm": 0.93359375, + "learning_rate": 6.537878559251824e-05, + "loss": 0.12739023, + "memory(GiB)": 58.24, + "step": 20360, + "train_speed(iter/s)": 0.767227 + }, + { + "epoch": 1.8975791433891993, + "grad_norm": 1.09375, + "learning_rate": 6.518628100982406e-05, + "loss": 0.1451817, + "memory(GiB)": 58.24, + "step": 20380, + "train_speed(iter/s)": 0.767254 + }, + { + "epoch": 1.899441340782123, + "grad_norm": 1.140625, + "learning_rate": 6.499392312091707e-05, + "loss": 0.14085768, + "memory(GiB)": 58.24, + "step": 20400, + "train_speed(iter/s)": 0.767302 + }, + { + "epoch": 1.9013035381750467, + "grad_norm": 1.1171875, + "learning_rate": 6.480171273633143e-05, + "loss": 0.14016054, + "memory(GiB)": 58.24, + "step": 20420, + "train_speed(iter/s)": 0.767337 + }, + { + "epoch": 1.90316573556797, + "grad_norm": 1.2265625, + "learning_rate": 6.46096506659797e-05, + "loss": 0.14577733, + "memory(GiB)": 58.24, + "step": 20440, + "train_speed(iter/s)": 0.767358 + }, + { + "epoch": 1.905027932960894, + "grad_norm": 1.296875, + "learning_rate": 6.441773771914947e-05, + "loss": 0.14134233, + "memory(GiB)": 58.24, + "step": 20460, + "train_speed(iter/s)": 0.767418 + }, + { + "epoch": 1.9068901303538175, + "grad_norm": 1.25, + "learning_rate": 6.422597470450014e-05, + "loss": 0.14478569, + "memory(GiB)": 58.24, + "step": 20480, + "train_speed(iter/s)": 0.767431 + }, + { + "epoch": 1.908752327746741, + "grad_norm": 1.1640625, + "learning_rate": 6.403436243005915e-05, + "loss": 0.12952732, + "memory(GiB)": 58.24, + "step": 20500, + "train_speed(iter/s)": 0.767457 + }, + { + "epoch": 1.910614525139665, + "grad_norm": 1.25, + "learning_rate": 6.384290170321881e-05, + "loss": 0.13815956, + "memory(GiB)": 58.24, + "step": 20520, + "train_speed(iter/s)": 0.767502 + }, + { + "epoch": 1.9124767225325885, + "grad_norm": 1.03125, + "learning_rate": 6.365159333073298e-05, + "loss": 0.14318891, + "memory(GiB)": 58.24, + "step": 20540, + "train_speed(iter/s)": 0.76757 + }, + { + "epoch": 1.914338919925512, + "grad_norm": 1.109375, + "learning_rate": 6.34604381187134e-05, + "loss": 0.13616923, + "memory(GiB)": 58.24, + "step": 20560, + "train_speed(iter/s)": 0.767628 + }, + { + "epoch": 1.916201117318436, + "grad_norm": 1.4140625, + "learning_rate": 6.326943687262656e-05, + "loss": 0.12770704, + "memory(GiB)": 58.24, + "step": 20580, + "train_speed(iter/s)": 0.767647 + }, + { + "epoch": 1.9180633147113593, + "grad_norm": 1.3359375, + "learning_rate": 6.307859039729007e-05, + "loss": 0.12648449, + "memory(GiB)": 58.24, + "step": 20600, + "train_speed(iter/s)": 0.767683 + }, + { + "epoch": 1.919925512104283, + "grad_norm": 1.515625, + "learning_rate": 6.288789949686941e-05, + "loss": 0.13040155, + "memory(GiB)": 58.24, + "step": 20620, + "train_speed(iter/s)": 0.767726 + }, + { + "epoch": 1.9217877094972067, + "grad_norm": 1.1484375, + "learning_rate": 6.269736497487469e-05, + "loss": 0.13270816, + "memory(GiB)": 58.24, + "step": 20640, + "train_speed(iter/s)": 0.767769 + }, + { + "epoch": 1.9236499068901303, + "grad_norm": 1.109375, + "learning_rate": 6.250698763415683e-05, + "loss": 0.14179718, + "memory(GiB)": 58.24, + "step": 20660, + "train_speed(iter/s)": 0.767801 + }, + { + "epoch": 1.925512104283054, + "grad_norm": 0.9140625, + "learning_rate": 6.231676827690457e-05, + "loss": 0.12922909, + "memory(GiB)": 58.24, + "step": 20680, + "train_speed(iter/s)": 0.767832 + }, + { + "epoch": 1.9273743016759777, + "grad_norm": 0.87109375, + "learning_rate": 6.212670770464102e-05, + "loss": 0.11642365, + "memory(GiB)": 58.24, + "step": 20700, + "train_speed(iter/s)": 0.76785 + }, + { + "epoch": 1.9292364990689013, + "grad_norm": 1.2578125, + "learning_rate": 6.19368067182201e-05, + "loss": 0.14179863, + "memory(GiB)": 58.24, + "step": 20720, + "train_speed(iter/s)": 0.767879 + }, + { + "epoch": 1.931098696461825, + "grad_norm": 1.0078125, + "learning_rate": 6.174706611782336e-05, + "loss": 0.14035552, + "memory(GiB)": 58.24, + "step": 20740, + "train_speed(iter/s)": 0.76793 + }, + { + "epoch": 1.9329608938547485, + "grad_norm": 1.265625, + "learning_rate": 6.155748670295654e-05, + "loss": 0.13533392, + "memory(GiB)": 58.24, + "step": 20760, + "train_speed(iter/s)": 0.767974 + }, + { + "epoch": 1.9348230912476723, + "grad_norm": 1.3671875, + "learning_rate": 6.136806927244607e-05, + "loss": 0.14196831, + "memory(GiB)": 58.24, + "step": 20780, + "train_speed(iter/s)": 0.767996 + }, + { + "epoch": 1.9366852886405959, + "grad_norm": 1.0234375, + "learning_rate": 6.117881462443604e-05, + "loss": 0.11872478, + "memory(GiB)": 58.24, + "step": 20800, + "train_speed(iter/s)": 0.76802 + }, + { + "epoch": 1.9385474860335195, + "grad_norm": 1.2890625, + "learning_rate": 6.098972355638448e-05, + "loss": 0.12733064, + "memory(GiB)": 58.24, + "step": 20820, + "train_speed(iter/s)": 0.768081 + }, + { + "epoch": 1.9404096834264433, + "grad_norm": 1.2890625, + "learning_rate": 6.080079686506014e-05, + "loss": 0.12513154, + "memory(GiB)": 58.24, + "step": 20840, + "train_speed(iter/s)": 0.768129 + }, + { + "epoch": 1.9422718808193669, + "grad_norm": 1.2421875, + "learning_rate": 6.061203534653926e-05, + "loss": 0.14031601, + "memory(GiB)": 58.24, + "step": 20860, + "train_speed(iter/s)": 0.768162 + }, + { + "epoch": 1.9441340782122905, + "grad_norm": 1.1953125, + "learning_rate": 6.042343979620197e-05, + "loss": 0.14372622, + "memory(GiB)": 58.24, + "step": 20880, + "train_speed(iter/s)": 0.768224 + }, + { + "epoch": 1.9459962756052143, + "grad_norm": 1.1640625, + "learning_rate": 6.0235011008729134e-05, + "loss": 0.13222244, + "memory(GiB)": 58.24, + "step": 20900, + "train_speed(iter/s)": 0.768263 + }, + { + "epoch": 1.9478584729981379, + "grad_norm": 0.9765625, + "learning_rate": 6.0046749778098875e-05, + "loss": 0.11751642, + "memory(GiB)": 58.24, + "step": 20920, + "train_speed(iter/s)": 0.768299 + }, + { + "epoch": 1.9497206703910615, + "grad_norm": 1.21875, + "learning_rate": 5.985865689758337e-05, + "loss": 0.13005948, + "memory(GiB)": 58.24, + "step": 20940, + "train_speed(iter/s)": 0.768318 + }, + { + "epoch": 1.9515828677839853, + "grad_norm": 1.4609375, + "learning_rate": 5.967073315974534e-05, + "loss": 0.13441616, + "memory(GiB)": 58.24, + "step": 20960, + "train_speed(iter/s)": 0.768371 + }, + { + "epoch": 1.9534450651769086, + "grad_norm": 1.0703125, + "learning_rate": 5.9482979356434834e-05, + "loss": 0.12724516, + "memory(GiB)": 58.24, + "step": 20980, + "train_speed(iter/s)": 0.768402 + }, + { + "epoch": 1.9553072625698324, + "grad_norm": 1.0859375, + "learning_rate": 5.929539627878582e-05, + "loss": 0.13387883, + "memory(GiB)": 58.24, + "step": 21000, + "train_speed(iter/s)": 0.768462 + }, + { + "epoch": 1.957169459962756, + "grad_norm": 0.8984375, + "learning_rate": 5.910798471721295e-05, + "loss": 0.13106408, + "memory(GiB)": 58.24, + "step": 21020, + "train_speed(iter/s)": 0.768497 + }, + { + "epoch": 1.9590316573556796, + "grad_norm": 0.84375, + "learning_rate": 5.892074546140809e-05, + "loss": 0.11956599, + "memory(GiB)": 58.24, + "step": 21040, + "train_speed(iter/s)": 0.76853 + }, + { + "epoch": 1.9608938547486034, + "grad_norm": 1.015625, + "learning_rate": 5.87336793003371e-05, + "loss": 0.12412496, + "memory(GiB)": 58.24, + "step": 21060, + "train_speed(iter/s)": 0.768568 + }, + { + "epoch": 1.962756052141527, + "grad_norm": 0.81640625, + "learning_rate": 5.854678702223648e-05, + "loss": 0.1311347, + "memory(GiB)": 58.24, + "step": 21080, + "train_speed(iter/s)": 0.768605 + }, + { + "epoch": 1.9646182495344506, + "grad_norm": 1.09375, + "learning_rate": 5.8360069414610016e-05, + "loss": 0.11528455, + "memory(GiB)": 58.24, + "step": 21100, + "train_speed(iter/s)": 0.768642 + }, + { + "epoch": 1.9664804469273744, + "grad_norm": 1.078125, + "learning_rate": 5.817352726422551e-05, + "loss": 0.14151897, + "memory(GiB)": 58.24, + "step": 21120, + "train_speed(iter/s)": 0.76868 + }, + { + "epoch": 1.9683426443202978, + "grad_norm": 1.1640625, + "learning_rate": 5.798716135711149e-05, + "loss": 0.13168578, + "memory(GiB)": 58.24, + "step": 21140, + "train_speed(iter/s)": 0.768704 + }, + { + "epoch": 1.9702048417132216, + "grad_norm": 0.82421875, + "learning_rate": 5.780097247855371e-05, + "loss": 0.11186194, + "memory(GiB)": 58.24, + "step": 21160, + "train_speed(iter/s)": 0.768737 + }, + { + "epoch": 1.9720670391061452, + "grad_norm": 1.3125, + "learning_rate": 5.76149614130921e-05, + "loss": 0.13154147, + "memory(GiB)": 58.24, + "step": 21180, + "train_speed(iter/s)": 0.768775 + }, + { + "epoch": 1.9739292364990688, + "grad_norm": 1.2578125, + "learning_rate": 5.742912894451735e-05, + "loss": 0.1238625, + "memory(GiB)": 58.24, + "step": 21200, + "train_speed(iter/s)": 0.768828 + }, + { + "epoch": 1.9757914338919926, + "grad_norm": 0.984375, + "learning_rate": 5.724347585586755e-05, + "loss": 0.11985067, + "memory(GiB)": 58.24, + "step": 21220, + "train_speed(iter/s)": 0.768858 + }, + { + "epoch": 1.9776536312849162, + "grad_norm": 1.28125, + "learning_rate": 5.705800292942498e-05, + "loss": 0.12159057, + "memory(GiB)": 58.24, + "step": 21240, + "train_speed(iter/s)": 0.768883 + }, + { + "epoch": 1.9795158286778398, + "grad_norm": 1.1953125, + "learning_rate": 5.687271094671266e-05, + "loss": 0.1149127, + "memory(GiB)": 58.24, + "step": 21260, + "train_speed(iter/s)": 0.768853 + }, + { + "epoch": 1.9813780260707636, + "grad_norm": 1.015625, + "learning_rate": 5.668760068849141e-05, + "loss": 0.13043272, + "memory(GiB)": 58.24, + "step": 21280, + "train_speed(iter/s)": 0.768896 + }, + { + "epoch": 1.983240223463687, + "grad_norm": 1.5, + "learning_rate": 5.650267293475605e-05, + "loss": 0.11858026, + "memory(GiB)": 58.24, + "step": 21300, + "train_speed(iter/s)": 0.768961 + }, + { + "epoch": 1.9851024208566108, + "grad_norm": 1.0546875, + "learning_rate": 5.631792846473256e-05, + "loss": 0.11108406, + "memory(GiB)": 58.24, + "step": 21320, + "train_speed(iter/s)": 0.769007 + }, + { + "epoch": 1.9869646182495344, + "grad_norm": 0.76953125, + "learning_rate": 5.6133368056874616e-05, + "loss": 0.1135129, + "memory(GiB)": 58.24, + "step": 21340, + "train_speed(iter/s)": 0.769044 + }, + { + "epoch": 1.988826815642458, + "grad_norm": 1.1015625, + "learning_rate": 5.594899248886012e-05, + "loss": 0.12456424, + "memory(GiB)": 58.24, + "step": 21360, + "train_speed(iter/s)": 0.769086 + }, + { + "epoch": 1.9906890130353818, + "grad_norm": 0.98828125, + "learning_rate": 5.5764802537588465e-05, + "loss": 0.12759299, + "memory(GiB)": 58.24, + "step": 21380, + "train_speed(iter/s)": 0.769135 + }, + { + "epoch": 1.9925512104283054, + "grad_norm": 1.3046875, + "learning_rate": 5.5580798979176595e-05, + "loss": 0.12331834, + "memory(GiB)": 58.24, + "step": 21400, + "train_speed(iter/s)": 0.769187 + }, + { + "epoch": 1.994413407821229, + "grad_norm": 1.0, + "learning_rate": 5.53969825889562e-05, + "loss": 0.11813304, + "memory(GiB)": 58.24, + "step": 21420, + "train_speed(iter/s)": 0.769246 + }, + { + "epoch": 1.9962756052141528, + "grad_norm": 1.2890625, + "learning_rate": 5.5213354141470294e-05, + "loss": 0.11984769, + "memory(GiB)": 58.24, + "step": 21440, + "train_speed(iter/s)": 0.769291 + }, + { + "epoch": 1.9981378026070762, + "grad_norm": 1.1328125, + "learning_rate": 5.502991441046998e-05, + "loss": 0.11140869, + "memory(GiB)": 58.24, + "step": 21460, + "train_speed(iter/s)": 0.769304 + }, + { + "epoch": 2.0, + "grad_norm": 1.2734375, + "learning_rate": 5.484666416891109e-05, + "loss": 0.12352898, + "memory(GiB)": 58.24, + "step": 21480, + "train_speed(iter/s)": 0.769312 + }, + { + "epoch": 2.001862197392924, + "grad_norm": 1.0625, + "learning_rate": 5.466360418895107e-05, + "loss": 0.12280267, + "memory(GiB)": 58.24, + "step": 21500, + "train_speed(iter/s)": 0.769021 + }, + { + "epoch": 2.003724394785847, + "grad_norm": 0.8671875, + "learning_rate": 5.4480735241945635e-05, + "loss": 0.12903731, + "memory(GiB)": 58.24, + "step": 21520, + "train_speed(iter/s)": 0.769074 + }, + { + "epoch": 2.005586592178771, + "grad_norm": 1.140625, + "learning_rate": 5.429805809844562e-05, + "loss": 0.13171415, + "memory(GiB)": 58.24, + "step": 21540, + "train_speed(iter/s)": 0.769104 + }, + { + "epoch": 2.007448789571695, + "grad_norm": 0.85546875, + "learning_rate": 5.4115573528193605e-05, + "loss": 0.10338333, + "memory(GiB)": 58.24, + "step": 21560, + "train_speed(iter/s)": 0.769152 + }, + { + "epoch": 2.009310986964618, + "grad_norm": 1.1328125, + "learning_rate": 5.3933282300120624e-05, + "loss": 0.12971883, + "memory(GiB)": 58.24, + "step": 21580, + "train_speed(iter/s)": 0.769168 + }, + { + "epoch": 2.011173184357542, + "grad_norm": 1.0859375, + "learning_rate": 5.3751185182343324e-05, + "loss": 0.12211378, + "memory(GiB)": 58.24, + "step": 21600, + "train_speed(iter/s)": 0.769214 + }, + { + "epoch": 2.0130353817504654, + "grad_norm": 0.8359375, + "learning_rate": 5.356928294216008e-05, + "loss": 0.11860096, + "memory(GiB)": 58.24, + "step": 21620, + "train_speed(iter/s)": 0.769226 + }, + { + "epoch": 2.014897579143389, + "grad_norm": 1.140625, + "learning_rate": 5.338757634604836e-05, + "loss": 0.12931483, + "memory(GiB)": 58.24, + "step": 21640, + "train_speed(iter/s)": 0.769265 + }, + { + "epoch": 2.016759776536313, + "grad_norm": 1.390625, + "learning_rate": 5.320606615966117e-05, + "loss": 0.11598758, + "memory(GiB)": 58.24, + "step": 21660, + "train_speed(iter/s)": 0.769256 + }, + { + "epoch": 2.0186219739292364, + "grad_norm": 1.234375, + "learning_rate": 5.302475314782389e-05, + "loss": 0.12075157, + "memory(GiB)": 58.24, + "step": 21680, + "train_speed(iter/s)": 0.769291 + }, + { + "epoch": 2.02048417132216, + "grad_norm": 0.984375, + "learning_rate": 5.284363807453117e-05, + "loss": 0.12393267, + "memory(GiB)": 58.24, + "step": 21700, + "train_speed(iter/s)": 0.769313 + }, + { + "epoch": 2.022346368715084, + "grad_norm": 0.875, + "learning_rate": 5.266272170294342e-05, + "loss": 0.11194068, + "memory(GiB)": 58.24, + "step": 21720, + "train_speed(iter/s)": 0.76932 + }, + { + "epoch": 2.0242085661080074, + "grad_norm": 1.015625, + "learning_rate": 5.248200479538394e-05, + "loss": 0.11720626, + "memory(GiB)": 58.24, + "step": 21740, + "train_speed(iter/s)": 0.769353 + }, + { + "epoch": 2.026070763500931, + "grad_norm": 0.875, + "learning_rate": 5.230148811333555e-05, + "loss": 0.10890026, + "memory(GiB)": 58.24, + "step": 21760, + "train_speed(iter/s)": 0.769387 + }, + { + "epoch": 2.0279329608938546, + "grad_norm": 1.0625, + "learning_rate": 5.212117241743735e-05, + "loss": 0.11610082, + "memory(GiB)": 58.24, + "step": 21780, + "train_speed(iter/s)": 0.769434 + }, + { + "epoch": 2.0297951582867784, + "grad_norm": 0.96875, + "learning_rate": 5.194105846748143e-05, + "loss": 0.13138523, + "memory(GiB)": 58.24, + "step": 21800, + "train_speed(iter/s)": 0.769494 + }, + { + "epoch": 2.031657355679702, + "grad_norm": 1.0078125, + "learning_rate": 5.176114702241006e-05, + "loss": 0.12315972, + "memory(GiB)": 58.24, + "step": 21820, + "train_speed(iter/s)": 0.76951 + }, + { + "epoch": 2.0335195530726256, + "grad_norm": 0.9765625, + "learning_rate": 5.158143884031197e-05, + "loss": 0.11700615, + "memory(GiB)": 58.24, + "step": 21840, + "train_speed(iter/s)": 0.769544 + }, + { + "epoch": 2.0353817504655494, + "grad_norm": 1.3203125, + "learning_rate": 5.1401934678419506e-05, + "loss": 0.12096615, + "memory(GiB)": 58.24, + "step": 21860, + "train_speed(iter/s)": 0.769577 + }, + { + "epoch": 2.037243947858473, + "grad_norm": 1.1640625, + "learning_rate": 5.122263529310538e-05, + "loss": 0.122633, + "memory(GiB)": 58.24, + "step": 21880, + "train_speed(iter/s)": 0.769619 + }, + { + "epoch": 2.0391061452513966, + "grad_norm": 1.0078125, + "learning_rate": 5.1043541439879286e-05, + "loss": 0.11288965, + "memory(GiB)": 58.24, + "step": 21900, + "train_speed(iter/s)": 0.769682 + }, + { + "epoch": 2.0409683426443204, + "grad_norm": 1.1171875, + "learning_rate": 5.0864653873385124e-05, + "loss": 0.12267628, + "memory(GiB)": 58.24, + "step": 21920, + "train_speed(iter/s)": 0.769729 + }, + { + "epoch": 2.0428305400372437, + "grad_norm": 1.4453125, + "learning_rate": 5.068597334739731e-05, + "loss": 0.11504929, + "memory(GiB)": 58.24, + "step": 21940, + "train_speed(iter/s)": 0.769754 + }, + { + "epoch": 2.0446927374301676, + "grad_norm": 1.203125, + "learning_rate": 5.0507500614817995e-05, + "loss": 0.11222873, + "memory(GiB)": 58.24, + "step": 21960, + "train_speed(iter/s)": 0.769798 + }, + { + "epoch": 2.0465549348230914, + "grad_norm": 0.84765625, + "learning_rate": 5.032923642767374e-05, + "loss": 0.11789845, + "memory(GiB)": 58.24, + "step": 21980, + "train_speed(iter/s)": 0.769844 + }, + { + "epoch": 2.0484171322160147, + "grad_norm": 1.140625, + "learning_rate": 5.015118153711237e-05, + "loss": 0.13255039, + "memory(GiB)": 58.24, + "step": 22000, + "train_speed(iter/s)": 0.769864 + }, + { + "epoch": 2.0484171322160147, + "eval_loss": 0.377461701631546, + "eval_runtime": 77.0906, + "eval_samples_per_second": 180.113, + "eval_steps_per_second": 1.414, + "step": 22000 + }, + { + "epoch": 2.0502793296089385, + "grad_norm": 1.0703125, + "learning_rate": 4.9973336693399795e-05, + "loss": 0.10642745, + "memory(GiB)": 58.24, + "step": 22020, + "train_speed(iter/s)": 0.766966 + }, + { + "epoch": 2.0521415270018624, + "grad_norm": 1.1640625, + "learning_rate": 4.97957026459168e-05, + "loss": 0.11735952, + "memory(GiB)": 58.24, + "step": 22040, + "train_speed(iter/s)": 0.766974 + }, + { + "epoch": 2.0540037243947857, + "grad_norm": 1.0234375, + "learning_rate": 4.961828014315599e-05, + "loss": 0.11628766, + "memory(GiB)": 58.24, + "step": 22060, + "train_speed(iter/s)": 0.766999 + }, + { + "epoch": 2.0558659217877095, + "grad_norm": 0.8359375, + "learning_rate": 4.944106993271863e-05, + "loss": 0.13654802, + "memory(GiB)": 58.24, + "step": 22080, + "train_speed(iter/s)": 0.767032 + }, + { + "epoch": 2.0577281191806334, + "grad_norm": 1.3515625, + "learning_rate": 4.926407276131141e-05, + "loss": 0.13798956, + "memory(GiB)": 58.24, + "step": 22100, + "train_speed(iter/s)": 0.767046 + }, + { + "epoch": 2.0595903165735567, + "grad_norm": 1.03125, + "learning_rate": 4.908728937474325e-05, + "loss": 0.10691235, + "memory(GiB)": 58.24, + "step": 22120, + "train_speed(iter/s)": 0.767074 + }, + { + "epoch": 2.0614525139664805, + "grad_norm": 0.9609375, + "learning_rate": 4.891072051792249e-05, + "loss": 0.11395538, + "memory(GiB)": 58.24, + "step": 22140, + "train_speed(iter/s)": 0.76709 + }, + { + "epoch": 2.063314711359404, + "grad_norm": 0.91796875, + "learning_rate": 4.873436693485325e-05, + "loss": 0.11044333, + "memory(GiB)": 58.24, + "step": 22160, + "train_speed(iter/s)": 0.767114 + }, + { + "epoch": 2.0651769087523277, + "grad_norm": 1.125, + "learning_rate": 4.8558229368632714e-05, + "loss": 0.11576631, + "memory(GiB)": 58.24, + "step": 22180, + "train_speed(iter/s)": 0.76713 + }, + { + "epoch": 2.0670391061452515, + "grad_norm": 1.171875, + "learning_rate": 4.838230856144781e-05, + "loss": 0.13391452, + "memory(GiB)": 58.24, + "step": 22200, + "train_speed(iter/s)": 0.767181 + }, + { + "epoch": 2.068901303538175, + "grad_norm": 1.2421875, + "learning_rate": 4.820660525457201e-05, + "loss": 0.10914934, + "memory(GiB)": 58.24, + "step": 22220, + "train_speed(iter/s)": 0.76721 + }, + { + "epoch": 2.0707635009310987, + "grad_norm": 1.1796875, + "learning_rate": 4.8031120188362546e-05, + "loss": 0.12316871, + "memory(GiB)": 58.24, + "step": 22240, + "train_speed(iter/s)": 0.767259 + }, + { + "epoch": 2.0726256983240225, + "grad_norm": 1.296875, + "learning_rate": 4.785585410225678e-05, + "loss": 0.11936328, + "memory(GiB)": 58.24, + "step": 22260, + "train_speed(iter/s)": 0.767306 + }, + { + "epoch": 2.074487895716946, + "grad_norm": 1.46875, + "learning_rate": 4.768080773476955e-05, + "loss": 0.11963451, + "memory(GiB)": 58.24, + "step": 22280, + "train_speed(iter/s)": 0.767332 + }, + { + "epoch": 2.0763500931098697, + "grad_norm": 1.234375, + "learning_rate": 4.7505981823489785e-05, + "loss": 0.12376244, + "memory(GiB)": 58.24, + "step": 22300, + "train_speed(iter/s)": 0.767376 + }, + { + "epoch": 2.078212290502793, + "grad_norm": 1.2890625, + "learning_rate": 4.7331377105077537e-05, + "loss": 0.12010782, + "memory(GiB)": 58.24, + "step": 22320, + "train_speed(iter/s)": 0.76741 + }, + { + "epoch": 2.080074487895717, + "grad_norm": 1.234375, + "learning_rate": 4.71569943152608e-05, + "loss": 0.10957891, + "memory(GiB)": 58.24, + "step": 22340, + "train_speed(iter/s)": 0.767455 + }, + { + "epoch": 2.0819366852886407, + "grad_norm": 0.92578125, + "learning_rate": 4.698283418883237e-05, + "loss": 0.10824106, + "memory(GiB)": 58.24, + "step": 22360, + "train_speed(iter/s)": 0.76747 + }, + { + "epoch": 2.083798882681564, + "grad_norm": 1.1796875, + "learning_rate": 4.680889745964689e-05, + "loss": 0.11550634, + "memory(GiB)": 58.24, + "step": 22380, + "train_speed(iter/s)": 0.767518 + }, + { + "epoch": 2.085661080074488, + "grad_norm": 0.97265625, + "learning_rate": 4.6635184860617675e-05, + "loss": 0.12425991, + "memory(GiB)": 58.24, + "step": 22400, + "train_speed(iter/s)": 0.767554 + }, + { + "epoch": 2.0875232774674117, + "grad_norm": 1.234375, + "learning_rate": 4.646169712371364e-05, + "loss": 0.1120156, + "memory(GiB)": 58.24, + "step": 22420, + "train_speed(iter/s)": 0.767568 + }, + { + "epoch": 2.089385474860335, + "grad_norm": 1.0078125, + "learning_rate": 4.628843497995606e-05, + "loss": 0.10975211, + "memory(GiB)": 58.24, + "step": 22440, + "train_speed(iter/s)": 0.767614 + }, + { + "epoch": 2.091247672253259, + "grad_norm": 1.3515625, + "learning_rate": 4.611539915941588e-05, + "loss": 0.1215034, + "memory(GiB)": 58.24, + "step": 22460, + "train_speed(iter/s)": 0.767646 + }, + { + "epoch": 2.0931098696461823, + "grad_norm": 0.97265625, + "learning_rate": 4.594259039121017e-05, + "loss": 0.1136503, + "memory(GiB)": 58.24, + "step": 22480, + "train_speed(iter/s)": 0.767685 + }, + { + "epoch": 2.094972067039106, + "grad_norm": 0.93359375, + "learning_rate": 4.577000940349939e-05, + "loss": 0.11370624, + "memory(GiB)": 58.24, + "step": 22500, + "train_speed(iter/s)": 0.76774 + }, + { + "epoch": 2.09683426443203, + "grad_norm": 0.96875, + "learning_rate": 4.559765692348421e-05, + "loss": 0.11183515, + "memory(GiB)": 58.24, + "step": 22520, + "train_speed(iter/s)": 0.767789 + }, + { + "epoch": 2.0986964618249533, + "grad_norm": 0.85546875, + "learning_rate": 4.54255336774023e-05, + "loss": 0.11438358, + "memory(GiB)": 58.24, + "step": 22540, + "train_speed(iter/s)": 0.767817 + }, + { + "epoch": 2.100558659217877, + "grad_norm": 0.94140625, + "learning_rate": 4.525364039052568e-05, + "loss": 0.10709119, + "memory(GiB)": 58.24, + "step": 22560, + "train_speed(iter/s)": 0.767825 + }, + { + "epoch": 2.102420856610801, + "grad_norm": 1.46875, + "learning_rate": 4.508197778715711e-05, + "loss": 0.12675226, + "memory(GiB)": 58.24, + "step": 22580, + "train_speed(iter/s)": 0.767876 + }, + { + "epoch": 2.1042830540037243, + "grad_norm": 1.125, + "learning_rate": 4.4910546590627525e-05, + "loss": 0.11741807, + "memory(GiB)": 58.24, + "step": 22600, + "train_speed(iter/s)": 0.767894 + }, + { + "epoch": 2.106145251396648, + "grad_norm": 1.21875, + "learning_rate": 4.473934752329269e-05, + "loss": 0.12141221, + "memory(GiB)": 58.24, + "step": 22620, + "train_speed(iter/s)": 0.767948 + }, + { + "epoch": 2.1080074487895715, + "grad_norm": 1.0546875, + "learning_rate": 4.456838130653026e-05, + "loss": 0.11729333, + "memory(GiB)": 58.24, + "step": 22640, + "train_speed(iter/s)": 0.767948 + }, + { + "epoch": 2.1098696461824953, + "grad_norm": 0.70703125, + "learning_rate": 4.439764866073682e-05, + "loss": 0.1138922, + "memory(GiB)": 58.24, + "step": 22660, + "train_speed(iter/s)": 0.767955 + }, + { + "epoch": 2.111731843575419, + "grad_norm": 1.3671875, + "learning_rate": 4.4227150305324605e-05, + "loss": 0.11240854, + "memory(GiB)": 58.24, + "step": 22680, + "train_speed(iter/s)": 0.768003 + }, + { + "epoch": 2.1135940409683425, + "grad_norm": 1.0625, + "learning_rate": 4.4056886958718744e-05, + "loss": 0.10561801, + "memory(GiB)": 58.24, + "step": 22700, + "train_speed(iter/s)": 0.768013 + }, + { + "epoch": 2.1154562383612663, + "grad_norm": 1.0078125, + "learning_rate": 4.388685933835409e-05, + "loss": 0.11626101, + "memory(GiB)": 58.24, + "step": 22720, + "train_speed(iter/s)": 0.768045 + }, + { + "epoch": 2.11731843575419, + "grad_norm": 1.03125, + "learning_rate": 4.3717068160672244e-05, + "loss": 0.10951095, + "memory(GiB)": 58.24, + "step": 22740, + "train_speed(iter/s)": 0.768074 + }, + { + "epoch": 2.1191806331471135, + "grad_norm": 0.9609375, + "learning_rate": 4.3547514141118394e-05, + "loss": 0.09885264, + "memory(GiB)": 58.24, + "step": 22760, + "train_speed(iter/s)": 0.768116 + }, + { + "epoch": 2.1210428305400373, + "grad_norm": 0.8984375, + "learning_rate": 4.337819799413864e-05, + "loss": 0.08843653, + "memory(GiB)": 58.24, + "step": 22780, + "train_speed(iter/s)": 0.768167 + }, + { + "epoch": 2.122905027932961, + "grad_norm": 0.9609375, + "learning_rate": 4.320912043317652e-05, + "loss": 0.10544789, + "memory(GiB)": 58.24, + "step": 22800, + "train_speed(iter/s)": 0.768184 + }, + { + "epoch": 2.1247672253258845, + "grad_norm": 0.9921875, + "learning_rate": 4.304028217067041e-05, + "loss": 0.10042415, + "memory(GiB)": 58.24, + "step": 22820, + "train_speed(iter/s)": 0.768232 + }, + { + "epoch": 2.1266294227188083, + "grad_norm": 1.328125, + "learning_rate": 4.287168391805031e-05, + "loss": 0.11463039, + "memory(GiB)": 58.24, + "step": 22840, + "train_speed(iter/s)": 0.768245 + }, + { + "epoch": 2.1284916201117317, + "grad_norm": 0.984375, + "learning_rate": 4.27033263857349e-05, + "loss": 0.11318433, + "memory(GiB)": 58.24, + "step": 22860, + "train_speed(iter/s)": 0.768273 + }, + { + "epoch": 2.1303538175046555, + "grad_norm": 1.0625, + "learning_rate": 4.253521028312857e-05, + "loss": 0.10956602, + "memory(GiB)": 58.24, + "step": 22880, + "train_speed(iter/s)": 0.768307 + }, + { + "epoch": 2.1322160148975793, + "grad_norm": 0.94921875, + "learning_rate": 4.23673363186183e-05, + "loss": 0.10085559, + "memory(GiB)": 58.24, + "step": 22900, + "train_speed(iter/s)": 0.768345 + }, + { + "epoch": 2.1340782122905027, + "grad_norm": 1.1484375, + "learning_rate": 4.2199705199570885e-05, + "loss": 0.10715935, + "memory(GiB)": 58.24, + "step": 22920, + "train_speed(iter/s)": 0.768374 + }, + { + "epoch": 2.1359404096834265, + "grad_norm": 1.1953125, + "learning_rate": 4.203231763232982e-05, + "loss": 0.12393291, + "memory(GiB)": 58.24, + "step": 22940, + "train_speed(iter/s)": 0.768377 + }, + { + "epoch": 2.1378026070763503, + "grad_norm": 1.125, + "learning_rate": 4.186517432221234e-05, + "loss": 0.10514045, + "memory(GiB)": 58.24, + "step": 22960, + "train_speed(iter/s)": 0.768416 + }, + { + "epoch": 2.1396648044692737, + "grad_norm": 1.4140625, + "learning_rate": 4.1698275973506464e-05, + "loss": 0.12016675, + "memory(GiB)": 58.24, + "step": 22980, + "train_speed(iter/s)": 0.768455 + }, + { + "epoch": 2.1415270018621975, + "grad_norm": 0.98046875, + "learning_rate": 4.153162328946806e-05, + "loss": 0.10668141, + "memory(GiB)": 58.24, + "step": 23000, + "train_speed(iter/s)": 0.76848 + }, + { + "epoch": 2.143389199255121, + "grad_norm": 0.921875, + "learning_rate": 4.136521697231773e-05, + "loss": 0.1124084, + "memory(GiB)": 58.24, + "step": 23020, + "train_speed(iter/s)": 0.768523 + }, + { + "epoch": 2.1452513966480447, + "grad_norm": 1.2265625, + "learning_rate": 4.1199057723238085e-05, + "loss": 0.11913981, + "memory(GiB)": 58.24, + "step": 23040, + "train_speed(iter/s)": 0.768557 + }, + { + "epoch": 2.1471135940409685, + "grad_norm": 0.9296875, + "learning_rate": 4.1033146242370656e-05, + "loss": 0.1068146, + "memory(GiB)": 58.24, + "step": 23060, + "train_speed(iter/s)": 0.768601 + }, + { + "epoch": 2.148975791433892, + "grad_norm": 0.8046875, + "learning_rate": 4.086748322881283e-05, + "loss": 0.10721408, + "memory(GiB)": 58.24, + "step": 23080, + "train_speed(iter/s)": 0.76861 + }, + { + "epoch": 2.1508379888268156, + "grad_norm": 1.0078125, + "learning_rate": 4.070206938061527e-05, + "loss": 0.11343349, + "memory(GiB)": 58.24, + "step": 23100, + "train_speed(iter/s)": 0.768649 + }, + { + "epoch": 2.1527001862197395, + "grad_norm": 1.1328125, + "learning_rate": 4.053690539477847e-05, + "loss": 0.11392477, + "memory(GiB)": 58.24, + "step": 23120, + "train_speed(iter/s)": 0.768699 + }, + { + "epoch": 2.154562383612663, + "grad_norm": 1.0625, + "learning_rate": 4.0371991967250336e-05, + "loss": 0.11984729, + "memory(GiB)": 58.24, + "step": 23140, + "train_speed(iter/s)": 0.768728 + }, + { + "epoch": 2.1564245810055866, + "grad_norm": 1.015625, + "learning_rate": 4.020732979292278e-05, + "loss": 0.12230155, + "memory(GiB)": 58.24, + "step": 23160, + "train_speed(iter/s)": 0.768733 + }, + { + "epoch": 2.1582867783985105, + "grad_norm": 0.9609375, + "learning_rate": 4.0042919565629155e-05, + "loss": 0.11182994, + "memory(GiB)": 58.24, + "step": 23180, + "train_speed(iter/s)": 0.768712 + }, + { + "epoch": 2.160148975791434, + "grad_norm": 0.91796875, + "learning_rate": 3.987876197814119e-05, + "loss": 0.1028029, + "memory(GiB)": 58.24, + "step": 23200, + "train_speed(iter/s)": 0.768722 + }, + { + "epoch": 2.1620111731843576, + "grad_norm": 0.96484375, + "learning_rate": 3.9714857722165956e-05, + "loss": 0.1047961, + "memory(GiB)": 58.24, + "step": 23220, + "train_speed(iter/s)": 0.768762 + }, + { + "epoch": 2.163873370577281, + "grad_norm": 1.8203125, + "learning_rate": 3.9551207488343165e-05, + "loss": 0.10931444, + "memory(GiB)": 58.24, + "step": 23240, + "train_speed(iter/s)": 0.768788 + }, + { + "epoch": 2.165735567970205, + "grad_norm": 1.0625, + "learning_rate": 3.938781196624215e-05, + "loss": 0.11954699, + "memory(GiB)": 58.24, + "step": 23260, + "train_speed(iter/s)": 0.768814 + }, + { + "epoch": 2.1675977653631286, + "grad_norm": 1.171875, + "learning_rate": 3.9224671844358964e-05, + "loss": 0.12002516, + "memory(GiB)": 58.24, + "step": 23280, + "train_speed(iter/s)": 0.768837 + }, + { + "epoch": 2.169459962756052, + "grad_norm": 1.21875, + "learning_rate": 3.906178781011347e-05, + "loss": 0.11104758, + "memory(GiB)": 58.24, + "step": 23300, + "train_speed(iter/s)": 0.768871 + }, + { + "epoch": 2.171322160148976, + "grad_norm": 1.25, + "learning_rate": 3.8899160549846504e-05, + "loss": 0.10640566, + "memory(GiB)": 58.24, + "step": 23320, + "train_speed(iter/s)": 0.768908 + }, + { + "epoch": 2.1731843575418996, + "grad_norm": 1.1171875, + "learning_rate": 3.873679074881685e-05, + "loss": 0.09920184, + "memory(GiB)": 58.24, + "step": 23340, + "train_speed(iter/s)": 0.768945 + }, + { + "epoch": 2.175046554934823, + "grad_norm": 1.0234375, + "learning_rate": 3.857467909119853e-05, + "loss": 0.09962378, + "memory(GiB)": 58.24, + "step": 23360, + "train_speed(iter/s)": 0.768952 + }, + { + "epoch": 2.176908752327747, + "grad_norm": 1.0078125, + "learning_rate": 3.841282626007784e-05, + "loss": 0.10315701, + "memory(GiB)": 74.75, + "step": 23380, + "train_speed(iter/s)": 0.768953 + }, + { + "epoch": 2.17877094972067, + "grad_norm": 1.09375, + "learning_rate": 3.82512329374503e-05, + "loss": 0.10988283, + "memory(GiB)": 74.75, + "step": 23400, + "train_speed(iter/s)": 0.768963 + }, + { + "epoch": 2.180633147113594, + "grad_norm": 1.0, + "learning_rate": 3.808989980421823e-05, + "loss": 0.10468407, + "memory(GiB)": 74.75, + "step": 23420, + "train_speed(iter/s)": 0.768989 + }, + { + "epoch": 2.182495344506518, + "grad_norm": 1.4609375, + "learning_rate": 3.7928827540187296e-05, + "loss": 0.1051132, + "memory(GiB)": 74.75, + "step": 23440, + "train_speed(iter/s)": 0.769015 + }, + { + "epoch": 2.184357541899441, + "grad_norm": 0.953125, + "learning_rate": 3.776801682406421e-05, + "loss": 0.11219141, + "memory(GiB)": 74.75, + "step": 23460, + "train_speed(iter/s)": 0.769046 + }, + { + "epoch": 2.186219739292365, + "grad_norm": 1.1796875, + "learning_rate": 3.7607468333453386e-05, + "loss": 0.09865897, + "memory(GiB)": 74.75, + "step": 23480, + "train_speed(iter/s)": 0.769062 + }, + { + "epoch": 2.188081936685289, + "grad_norm": 0.98046875, + "learning_rate": 3.744718274485445e-05, + "loss": 0.10719773, + "memory(GiB)": 74.75, + "step": 23500, + "train_speed(iter/s)": 0.769096 + }, + { + "epoch": 2.189944134078212, + "grad_norm": 0.921875, + "learning_rate": 3.728716073365921e-05, + "loss": 0.10016888, + "memory(GiB)": 74.75, + "step": 23520, + "train_speed(iter/s)": 0.769102 + }, + { + "epoch": 2.191806331471136, + "grad_norm": 0.90234375, + "learning_rate": 3.712740297414875e-05, + "loss": 0.09913735, + "memory(GiB)": 74.75, + "step": 23540, + "train_speed(iter/s)": 0.769151 + }, + { + "epoch": 2.1936685288640594, + "grad_norm": 0.7578125, + "learning_rate": 3.696791013949081e-05, + "loss": 0.11304789, + "memory(GiB)": 74.75, + "step": 23560, + "train_speed(iter/s)": 0.7692 + }, + { + "epoch": 2.195530726256983, + "grad_norm": 0.97265625, + "learning_rate": 3.680868290173677e-05, + "loss": 0.11336186, + "memory(GiB)": 74.75, + "step": 23580, + "train_speed(iter/s)": 0.769216 + }, + { + "epoch": 2.197392923649907, + "grad_norm": 1.296875, + "learning_rate": 3.664972193181887e-05, + "loss": 0.09854193, + "memory(GiB)": 74.75, + "step": 23600, + "train_speed(iter/s)": 0.769212 + }, + { + "epoch": 2.1992551210428304, + "grad_norm": 0.90234375, + "learning_rate": 3.649102789954738e-05, + "loss": 0.09684347, + "memory(GiB)": 74.75, + "step": 23620, + "train_speed(iter/s)": 0.769264 + }, + { + "epoch": 2.201117318435754, + "grad_norm": 1.1328125, + "learning_rate": 3.633260147360783e-05, + "loss": 0.10960507, + "memory(GiB)": 74.75, + "step": 23640, + "train_speed(iter/s)": 0.769308 + }, + { + "epoch": 2.202979515828678, + "grad_norm": 1.078125, + "learning_rate": 3.617444332155803e-05, + "loss": 0.11710701, + "memory(GiB)": 74.75, + "step": 23660, + "train_speed(iter/s)": 0.769343 + }, + { + "epoch": 2.2048417132216014, + "grad_norm": 0.859375, + "learning_rate": 3.601655410982545e-05, + "loss": 0.10495639, + "memory(GiB)": 74.75, + "step": 23680, + "train_speed(iter/s)": 0.769394 + }, + { + "epoch": 2.206703910614525, + "grad_norm": 0.89453125, + "learning_rate": 3.585893450370439e-05, + "loss": 0.10962081, + "memory(GiB)": 74.75, + "step": 23700, + "train_speed(iter/s)": 0.769409 + }, + { + "epoch": 2.2085661080074486, + "grad_norm": 0.875, + "learning_rate": 3.5701585167352924e-05, + "loss": 0.11209228, + "memory(GiB)": 74.75, + "step": 23720, + "train_speed(iter/s)": 0.769435 + }, + { + "epoch": 2.2104283054003724, + "grad_norm": 1.3359375, + "learning_rate": 3.554450676379056e-05, + "loss": 0.10384083, + "memory(GiB)": 74.75, + "step": 23740, + "train_speed(iter/s)": 0.769455 + }, + { + "epoch": 2.212290502793296, + "grad_norm": 0.9375, + "learning_rate": 3.538769995489494e-05, + "loss": 0.10817424, + "memory(GiB)": 74.75, + "step": 23760, + "train_speed(iter/s)": 0.769461 + }, + { + "epoch": 2.2141527001862196, + "grad_norm": 0.84375, + "learning_rate": 3.523116540139949e-05, + "loss": 0.11751704, + "memory(GiB)": 74.75, + "step": 23780, + "train_speed(iter/s)": 0.769484 + }, + { + "epoch": 2.2160148975791434, + "grad_norm": 1.21875, + "learning_rate": 3.507490376289029e-05, + "loss": 0.11738909, + "memory(GiB)": 74.75, + "step": 23800, + "train_speed(iter/s)": 0.769518 + }, + { + "epoch": 2.217877094972067, + "grad_norm": 1.2109375, + "learning_rate": 3.49189156978035e-05, + "loss": 0.10404855, + "memory(GiB)": 74.75, + "step": 23820, + "train_speed(iter/s)": 0.769528 + }, + { + "epoch": 2.2197392923649906, + "grad_norm": 0.82421875, + "learning_rate": 3.476320186342259e-05, + "loss": 0.10161319, + "memory(GiB)": 74.75, + "step": 23840, + "train_speed(iter/s)": 0.76957 + }, + { + "epoch": 2.2216014897579144, + "grad_norm": 1.140625, + "learning_rate": 3.4607762915875374e-05, + "loss": 0.10832006, + "memory(GiB)": 74.75, + "step": 23860, + "train_speed(iter/s)": 0.7696 + }, + { + "epoch": 2.223463687150838, + "grad_norm": 1.34375, + "learning_rate": 3.445259951013149e-05, + "loss": 0.1030372, + "memory(GiB)": 74.75, + "step": 23880, + "train_speed(iter/s)": 0.76962 + }, + { + "epoch": 2.2253258845437616, + "grad_norm": 0.86328125, + "learning_rate": 3.4297712299999515e-05, + "loss": 0.09935099, + "memory(GiB)": 74.75, + "step": 23900, + "train_speed(iter/s)": 0.769668 + }, + { + "epoch": 2.2271880819366854, + "grad_norm": 1.515625, + "learning_rate": 3.4143101938124223e-05, + "loss": 0.11144433, + "memory(GiB)": 74.75, + "step": 23920, + "train_speed(iter/s)": 0.769691 + }, + { + "epoch": 2.2290502793296088, + "grad_norm": 0.99609375, + "learning_rate": 3.398876907598379e-05, + "loss": 0.11325089, + "memory(GiB)": 74.75, + "step": 23940, + "train_speed(iter/s)": 0.769745 + }, + { + "epoch": 2.2309124767225326, + "grad_norm": 1.0, + "learning_rate": 3.383471436388722e-05, + "loss": 0.10212806, + "memory(GiB)": 74.75, + "step": 23960, + "train_speed(iter/s)": 0.769766 + }, + { + "epoch": 2.2327746741154564, + "grad_norm": 1.078125, + "learning_rate": 3.36809384509713e-05, + "loss": 0.12155344, + "memory(GiB)": 74.75, + "step": 23980, + "train_speed(iter/s)": 0.769802 + }, + { + "epoch": 2.2346368715083798, + "grad_norm": 1.0625, + "learning_rate": 3.352744198519818e-05, + "loss": 0.1001812, + "memory(GiB)": 74.75, + "step": 24000, + "train_speed(iter/s)": 0.769804 + }, + { + "epoch": 2.2346368715083798, + "eval_loss": 0.38177013397216797, + "eval_runtime": 77.0804, + "eval_samples_per_second": 180.137, + "eval_steps_per_second": 1.414, + "step": 24000 + }, + { + "epoch": 2.2364990689013036, + "grad_norm": 1.140625, + "learning_rate": 3.337422561335249e-05, + "loss": 0.10155301, + "memory(GiB)": 74.75, + "step": 24020, + "train_speed(iter/s)": 0.767136 + }, + { + "epoch": 2.2383612662942274, + "grad_norm": 0.84765625, + "learning_rate": 3.322128998103863e-05, + "loss": 0.10558331, + "memory(GiB)": 74.75, + "step": 24040, + "train_speed(iter/s)": 0.767149 + }, + { + "epoch": 2.2402234636871508, + "grad_norm": 1.2421875, + "learning_rate": 3.3068635732678044e-05, + "loss": 0.10001249, + "memory(GiB)": 74.75, + "step": 24060, + "train_speed(iter/s)": 0.767192 + }, + { + "epoch": 2.2420856610800746, + "grad_norm": 1.0859375, + "learning_rate": 3.2916263511506453e-05, + "loss": 0.10083565, + "memory(GiB)": 74.75, + "step": 24080, + "train_speed(iter/s)": 0.767228 + }, + { + "epoch": 2.243947858472998, + "grad_norm": 1.0, + "learning_rate": 3.276417395957138e-05, + "loss": 0.11419288, + "memory(GiB)": 74.75, + "step": 24100, + "train_speed(iter/s)": 0.767252 + }, + { + "epoch": 2.2458100558659218, + "grad_norm": 1.2578125, + "learning_rate": 3.2612367717729054e-05, + "loss": 0.10783464, + "memory(GiB)": 74.75, + "step": 24120, + "train_speed(iter/s)": 0.767284 + }, + { + "epoch": 2.2476722532588456, + "grad_norm": 0.9453125, + "learning_rate": 3.2460845425642095e-05, + "loss": 0.11875585, + "memory(GiB)": 74.75, + "step": 24140, + "train_speed(iter/s)": 0.767304 + }, + { + "epoch": 2.249534450651769, + "grad_norm": 1.375, + "learning_rate": 3.230960772177656e-05, + "loss": 0.09926898, + "memory(GiB)": 74.75, + "step": 24160, + "train_speed(iter/s)": 0.767332 + }, + { + "epoch": 2.2513966480446927, + "grad_norm": 1.0234375, + "learning_rate": 3.215865524339943e-05, + "loss": 0.09872712, + "memory(GiB)": 74.75, + "step": 24180, + "train_speed(iter/s)": 0.767362 + }, + { + "epoch": 2.2532588454376166, + "grad_norm": 1.3046875, + "learning_rate": 3.200798862657568e-05, + "loss": 0.10884237, + "memory(GiB)": 74.75, + "step": 24200, + "train_speed(iter/s)": 0.767404 + }, + { + "epoch": 2.25512104283054, + "grad_norm": 1.0390625, + "learning_rate": 3.185760850616592e-05, + "loss": 0.11168021, + "memory(GiB)": 74.75, + "step": 24220, + "train_speed(iter/s)": 0.767435 + }, + { + "epoch": 2.2569832402234637, + "grad_norm": 0.76953125, + "learning_rate": 3.170751551582347e-05, + "loss": 0.10452518, + "memory(GiB)": 74.75, + "step": 24240, + "train_speed(iter/s)": 0.767454 + }, + { + "epoch": 2.2588454376163876, + "grad_norm": 0.8359375, + "learning_rate": 3.155771028799182e-05, + "loss": 0.10053955, + "memory(GiB)": 74.75, + "step": 24260, + "train_speed(iter/s)": 0.767471 + }, + { + "epoch": 2.260707635009311, + "grad_norm": 1.109375, + "learning_rate": 3.140819345390191e-05, + "loss": 0.11290877, + "memory(GiB)": 74.75, + "step": 24280, + "train_speed(iter/s)": 0.767523 + }, + { + "epoch": 2.2625698324022347, + "grad_norm": 1.046875, + "learning_rate": 3.125896564356938e-05, + "loss": 0.0997721, + "memory(GiB)": 74.75, + "step": 24300, + "train_speed(iter/s)": 0.767559 + }, + { + "epoch": 2.264432029795158, + "grad_norm": 0.7265625, + "learning_rate": 3.111002748579226e-05, + "loss": 0.10257297, + "memory(GiB)": 74.75, + "step": 24320, + "train_speed(iter/s)": 0.767592 + }, + { + "epoch": 2.266294227188082, + "grad_norm": 1.15625, + "learning_rate": 3.09613796081478e-05, + "loss": 0.1102082, + "memory(GiB)": 74.75, + "step": 24340, + "train_speed(iter/s)": 0.767628 + }, + { + "epoch": 2.2681564245810057, + "grad_norm": 0.9765625, + "learning_rate": 3.0813022636990275e-05, + "loss": 0.11674392, + "memory(GiB)": 74.75, + "step": 24360, + "train_speed(iter/s)": 0.767675 + }, + { + "epoch": 2.270018621973929, + "grad_norm": 1.25, + "learning_rate": 3.066495719744815e-05, + "loss": 0.09853878, + "memory(GiB)": 74.75, + "step": 24380, + "train_speed(iter/s)": 0.767684 + }, + { + "epoch": 2.271880819366853, + "grad_norm": 0.83984375, + "learning_rate": 3.0517183913421367e-05, + "loss": 0.1074196, + "memory(GiB)": 74.75, + "step": 24400, + "train_speed(iter/s)": 0.767675 + }, + { + "epoch": 2.2737430167597763, + "grad_norm": 1.453125, + "learning_rate": 3.036970340757903e-05, + "loss": 0.09881254, + "memory(GiB)": 74.75, + "step": 24420, + "train_speed(iter/s)": 0.767677 + }, + { + "epoch": 2.2756052141527, + "grad_norm": 1.1171875, + "learning_rate": 3.022251630135633e-05, + "loss": 0.11979437, + "memory(GiB)": 74.75, + "step": 24440, + "train_speed(iter/s)": 0.767719 + }, + { + "epoch": 2.277467411545624, + "grad_norm": 1.4609375, + "learning_rate": 3.007562321495231e-05, + "loss": 0.10171646, + "memory(GiB)": 74.75, + "step": 24460, + "train_speed(iter/s)": 0.767738 + }, + { + "epoch": 2.2793296089385473, + "grad_norm": 1.203125, + "learning_rate": 2.9929024767327086e-05, + "loss": 0.10340936, + "memory(GiB)": 74.75, + "step": 24480, + "train_speed(iter/s)": 0.767767 + }, + { + "epoch": 2.281191806331471, + "grad_norm": 1.1953125, + "learning_rate": 2.9782721576199268e-05, + "loss": 0.09857596, + "memory(GiB)": 74.75, + "step": 24500, + "train_speed(iter/s)": 0.767785 + }, + { + "epoch": 2.283054003724395, + "grad_norm": 1.0546875, + "learning_rate": 2.9636714258043273e-05, + "loss": 0.10468978, + "memory(GiB)": 74.75, + "step": 24520, + "train_speed(iter/s)": 0.767803 + }, + { + "epoch": 2.2849162011173183, + "grad_norm": 1.046875, + "learning_rate": 2.9491003428086905e-05, + "loss": 0.10225184, + "memory(GiB)": 74.75, + "step": 24540, + "train_speed(iter/s)": 0.767846 + }, + { + "epoch": 2.286778398510242, + "grad_norm": 0.84765625, + "learning_rate": 2.934558970030862e-05, + "loss": 0.09291599, + "memory(GiB)": 74.75, + "step": 24560, + "train_speed(iter/s)": 0.767871 + }, + { + "epoch": 2.288640595903166, + "grad_norm": 1.078125, + "learning_rate": 2.920047368743497e-05, + "loss": 0.102846, + "memory(GiB)": 74.75, + "step": 24580, + "train_speed(iter/s)": 0.7679 + }, + { + "epoch": 2.2905027932960893, + "grad_norm": 1.1015625, + "learning_rate": 2.9055656000938092e-05, + "loss": 0.09924616, + "memory(GiB)": 74.75, + "step": 24600, + "train_speed(iter/s)": 0.767908 + }, + { + "epoch": 2.292364990689013, + "grad_norm": 1.4921875, + "learning_rate": 2.8911137251032915e-05, + "loss": 0.11189671, + "memory(GiB)": 74.75, + "step": 24620, + "train_speed(iter/s)": 0.767926 + }, + { + "epoch": 2.294227188081937, + "grad_norm": 1.0078125, + "learning_rate": 2.8766918046674995e-05, + "loss": 0.10130365, + "memory(GiB)": 74.75, + "step": 24640, + "train_speed(iter/s)": 0.767934 + }, + { + "epoch": 2.2960893854748603, + "grad_norm": 1.359375, + "learning_rate": 2.862299899555746e-05, + "loss": 0.10901361, + "memory(GiB)": 74.75, + "step": 24660, + "train_speed(iter/s)": 0.767971 + }, + { + "epoch": 2.297951582867784, + "grad_norm": 0.875, + "learning_rate": 2.8479380704108815e-05, + "loss": 0.10119644, + "memory(GiB)": 74.75, + "step": 24680, + "train_speed(iter/s)": 0.768011 + }, + { + "epoch": 2.2998137802607075, + "grad_norm": 0.9375, + "learning_rate": 2.8336063777490275e-05, + "loss": 0.10090758, + "memory(GiB)": 74.75, + "step": 24700, + "train_speed(iter/s)": 0.768031 + }, + { + "epoch": 2.3016759776536313, + "grad_norm": 1.3203125, + "learning_rate": 2.819304881959306e-05, + "loss": 0.11179039, + "memory(GiB)": 74.75, + "step": 24720, + "train_speed(iter/s)": 0.768069 + }, + { + "epoch": 2.303538175046555, + "grad_norm": 0.953125, + "learning_rate": 2.8050336433036216e-05, + "loss": 0.09195191, + "memory(GiB)": 74.75, + "step": 24740, + "train_speed(iter/s)": 0.768106 + }, + { + "epoch": 2.3054003724394785, + "grad_norm": 1.0, + "learning_rate": 2.7907927219163654e-05, + "loss": 0.09057497, + "memory(GiB)": 74.75, + "step": 24760, + "train_speed(iter/s)": 0.768137 + }, + { + "epoch": 2.3072625698324023, + "grad_norm": 1.0546875, + "learning_rate": 2.776582177804189e-05, + "loss": 0.10486867, + "memory(GiB)": 74.75, + "step": 24780, + "train_speed(iter/s)": 0.768152 + }, + { + "epoch": 2.3091247672253257, + "grad_norm": 0.98828125, + "learning_rate": 2.7624020708457442e-05, + "loss": 0.10587823, + "memory(GiB)": 74.75, + "step": 24800, + "train_speed(iter/s)": 0.768172 + }, + { + "epoch": 2.3109869646182495, + "grad_norm": 1.15625, + "learning_rate": 2.748252460791433e-05, + "loss": 0.10775185, + "memory(GiB)": 74.75, + "step": 24820, + "train_speed(iter/s)": 0.768207 + }, + { + "epoch": 2.3128491620111733, + "grad_norm": 1.1328125, + "learning_rate": 2.7341334072631454e-05, + "loss": 0.09222878, + "memory(GiB)": 74.75, + "step": 24840, + "train_speed(iter/s)": 0.768225 + }, + { + "epoch": 2.3147113594040967, + "grad_norm": 1.3203125, + "learning_rate": 2.720044969754021e-05, + "loss": 0.10134428, + "memory(GiB)": 74.75, + "step": 24860, + "train_speed(iter/s)": 0.768226 + }, + { + "epoch": 2.3165735567970205, + "grad_norm": 1.0859375, + "learning_rate": 2.7059872076281956e-05, + "loss": 0.09082563, + "memory(GiB)": 74.75, + "step": 24880, + "train_speed(iter/s)": 0.768223 + }, + { + "epoch": 2.3184357541899443, + "grad_norm": 0.95703125, + "learning_rate": 2.691960180120544e-05, + "loss": 0.10442271, + "memory(GiB)": 74.75, + "step": 24900, + "train_speed(iter/s)": 0.768261 + }, + { + "epoch": 2.3202979515828677, + "grad_norm": 1.171875, + "learning_rate": 2.6779639463364438e-05, + "loss": 0.1070719, + "memory(GiB)": 74.75, + "step": 24920, + "train_speed(iter/s)": 0.768305 + }, + { + "epoch": 2.3221601489757915, + "grad_norm": 1.0234375, + "learning_rate": 2.6639985652514987e-05, + "loss": 0.10250396, + "memory(GiB)": 74.75, + "step": 24940, + "train_speed(iter/s)": 0.768353 + }, + { + "epoch": 2.3240223463687153, + "grad_norm": 0.84765625, + "learning_rate": 2.6500640957113364e-05, + "loss": 0.09741715, + "memory(GiB)": 74.75, + "step": 24960, + "train_speed(iter/s)": 0.768386 + }, + { + "epoch": 2.3258845437616387, + "grad_norm": 1.1328125, + "learning_rate": 2.6361605964313084e-05, + "loss": 0.09976978, + "memory(GiB)": 74.75, + "step": 24980, + "train_speed(iter/s)": 0.768395 + }, + { + "epoch": 2.3277467411545625, + "grad_norm": 0.8828125, + "learning_rate": 2.6222881259962828e-05, + "loss": 0.09276388, + "memory(GiB)": 74.75, + "step": 25000, + "train_speed(iter/s)": 0.768413 + }, + { + "epoch": 2.329608938547486, + "grad_norm": 1.1953125, + "learning_rate": 2.6084467428603788e-05, + "loss": 0.10102336, + "memory(GiB)": 74.75, + "step": 25020, + "train_speed(iter/s)": 0.768435 + }, + { + "epoch": 2.3314711359404097, + "grad_norm": 1.0390625, + "learning_rate": 2.5946365053467126e-05, + "loss": 0.09177208, + "memory(GiB)": 74.75, + "step": 25040, + "train_speed(iter/s)": 0.768475 + }, + { + "epoch": 2.3333333333333335, + "grad_norm": 0.87890625, + "learning_rate": 2.5808574716471856e-05, + "loss": 0.10831335, + "memory(GiB)": 74.75, + "step": 25060, + "train_speed(iter/s)": 0.768501 + }, + { + "epoch": 2.335195530726257, + "grad_norm": 0.765625, + "learning_rate": 2.5671096998221943e-05, + "loss": 0.08765578, + "memory(GiB)": 74.75, + "step": 25080, + "train_speed(iter/s)": 0.768518 + }, + { + "epoch": 2.3370577281191807, + "grad_norm": 1.2578125, + "learning_rate": 2.5533932478004186e-05, + "loss": 0.10742819, + "memory(GiB)": 74.75, + "step": 25100, + "train_speed(iter/s)": 0.76853 + }, + { + "epoch": 2.338919925512104, + "grad_norm": 1.0546875, + "learning_rate": 2.5397081733785654e-05, + "loss": 0.09265711, + "memory(GiB)": 74.75, + "step": 25120, + "train_speed(iter/s)": 0.768552 + }, + { + "epoch": 2.340782122905028, + "grad_norm": 1.1328125, + "learning_rate": 2.5260545342211295e-05, + "loss": 0.11310081, + "memory(GiB)": 74.75, + "step": 25140, + "train_speed(iter/s)": 0.768573 + }, + { + "epoch": 2.3426443202979517, + "grad_norm": 1.1796875, + "learning_rate": 2.5124323878601342e-05, + "loss": 0.09951773, + "memory(GiB)": 74.75, + "step": 25160, + "train_speed(iter/s)": 0.768598 + }, + { + "epoch": 2.344506517690875, + "grad_norm": 0.96484375, + "learning_rate": 2.498841791694925e-05, + "loss": 0.0946602, + "memory(GiB)": 74.75, + "step": 25180, + "train_speed(iter/s)": 0.768608 + }, + { + "epoch": 2.346368715083799, + "grad_norm": 0.9375, + "learning_rate": 2.4852828029918817e-05, + "loss": 0.0985311, + "memory(GiB)": 74.75, + "step": 25200, + "train_speed(iter/s)": 0.768644 + }, + { + "epoch": 2.3482309124767227, + "grad_norm": 0.80078125, + "learning_rate": 2.4717554788842156e-05, + "loss": 0.10430787, + "memory(GiB)": 74.75, + "step": 25220, + "train_speed(iter/s)": 0.768664 + }, + { + "epoch": 2.350093109869646, + "grad_norm": 1.078125, + "learning_rate": 2.4582598763717125e-05, + "loss": 0.10131855, + "memory(GiB)": 74.75, + "step": 25240, + "train_speed(iter/s)": 0.768693 + }, + { + "epoch": 2.35195530726257, + "grad_norm": 0.88671875, + "learning_rate": 2.4447960523204795e-05, + "loss": 0.09341418, + "memory(GiB)": 74.75, + "step": 25260, + "train_speed(iter/s)": 0.768703 + }, + { + "epoch": 2.3538175046554937, + "grad_norm": 0.81640625, + "learning_rate": 2.431364063462743e-05, + "loss": 0.1005939, + "memory(GiB)": 74.75, + "step": 25280, + "train_speed(iter/s)": 0.768753 + }, + { + "epoch": 2.355679702048417, + "grad_norm": 1.015625, + "learning_rate": 2.417963966396565e-05, + "loss": 0.09922752, + "memory(GiB)": 74.75, + "step": 25300, + "train_speed(iter/s)": 0.76878 + }, + { + "epoch": 2.357541899441341, + "grad_norm": 1.140625, + "learning_rate": 2.4045958175856354e-05, + "loss": 0.10236137, + "memory(GiB)": 74.75, + "step": 25320, + "train_speed(iter/s)": 0.768813 + }, + { + "epoch": 2.3594040968342647, + "grad_norm": 0.99609375, + "learning_rate": 2.3912596733590242e-05, + "loss": 0.10635381, + "memory(GiB)": 74.75, + "step": 25340, + "train_speed(iter/s)": 0.768832 + }, + { + "epoch": 2.361266294227188, + "grad_norm": 0.953125, + "learning_rate": 2.377955589910943e-05, + "loss": 0.10511913, + "memory(GiB)": 74.75, + "step": 25360, + "train_speed(iter/s)": 0.768863 + }, + { + "epoch": 2.363128491620112, + "grad_norm": 0.91796875, + "learning_rate": 2.364683623300513e-05, + "loss": 0.08750724, + "memory(GiB)": 74.75, + "step": 25380, + "train_speed(iter/s)": 0.7689 + }, + { + "epoch": 2.364990689013035, + "grad_norm": 0.90625, + "learning_rate": 2.3514438294515172e-05, + "loss": 0.08906246, + "memory(GiB)": 74.75, + "step": 25400, + "train_speed(iter/s)": 0.768923 + }, + { + "epoch": 2.366852886405959, + "grad_norm": 1.0546875, + "learning_rate": 2.3382362641521803e-05, + "loss": 0.10484225, + "memory(GiB)": 74.75, + "step": 25420, + "train_speed(iter/s)": 0.768941 + }, + { + "epoch": 2.368715083798883, + "grad_norm": 0.921875, + "learning_rate": 2.325060983054924e-05, + "loss": 0.10104049, + "memory(GiB)": 74.75, + "step": 25440, + "train_speed(iter/s)": 0.768962 + }, + { + "epoch": 2.370577281191806, + "grad_norm": 0.9609375, + "learning_rate": 2.3119180416761378e-05, + "loss": 0.10175326, + "memory(GiB)": 74.75, + "step": 25460, + "train_speed(iter/s)": 0.768988 + }, + { + "epoch": 2.37243947858473, + "grad_norm": 1.21875, + "learning_rate": 2.2988074953959317e-05, + "loss": 0.096415, + "memory(GiB)": 74.75, + "step": 25480, + "train_speed(iter/s)": 0.769004 + }, + { + "epoch": 2.3743016759776534, + "grad_norm": 1.0859375, + "learning_rate": 2.28572939945793e-05, + "loss": 0.09378529, + "memory(GiB)": 74.75, + "step": 25500, + "train_speed(iter/s)": 0.768992 + }, + { + "epoch": 2.376163873370577, + "grad_norm": 1.03125, + "learning_rate": 2.272683808969004e-05, + "loss": 0.10414027, + "memory(GiB)": 74.75, + "step": 25520, + "train_speed(iter/s)": 0.769017 + }, + { + "epoch": 2.378026070763501, + "grad_norm": 0.87109375, + "learning_rate": 2.259670778899071e-05, + "loss": 0.10215082, + "memory(GiB)": 74.75, + "step": 25540, + "train_speed(iter/s)": 0.769021 + }, + { + "epoch": 2.3798882681564244, + "grad_norm": 0.984375, + "learning_rate": 2.2466903640808446e-05, + "loss": 0.09546254, + "memory(GiB)": 74.75, + "step": 25560, + "train_speed(iter/s)": 0.769032 + }, + { + "epoch": 2.381750465549348, + "grad_norm": 1.296875, + "learning_rate": 2.2337426192096002e-05, + "loss": 0.10957, + "memory(GiB)": 74.75, + "step": 25580, + "train_speed(iter/s)": 0.769066 + }, + { + "epoch": 2.383612662942272, + "grad_norm": 1.15625, + "learning_rate": 2.2208275988429716e-05, + "loss": 0.09794831, + "memory(GiB)": 74.75, + "step": 25600, + "train_speed(iter/s)": 0.769101 + }, + { + "epoch": 2.3854748603351954, + "grad_norm": 0.76953125, + "learning_rate": 2.2079453574006814e-05, + "loss": 0.09378864, + "memory(GiB)": 74.75, + "step": 25620, + "train_speed(iter/s)": 0.769125 + }, + { + "epoch": 2.387337057728119, + "grad_norm": 1.015625, + "learning_rate": 2.1950959491643454e-05, + "loss": 0.10166427, + "memory(GiB)": 74.75, + "step": 25640, + "train_speed(iter/s)": 0.769157 + }, + { + "epoch": 2.389199255121043, + "grad_norm": 0.78515625, + "learning_rate": 2.1822794282772253e-05, + "loss": 0.09407473, + "memory(GiB)": 74.75, + "step": 25660, + "train_speed(iter/s)": 0.769197 + }, + { + "epoch": 2.3910614525139664, + "grad_norm": 1.0625, + "learning_rate": 2.169495848744012e-05, + "loss": 0.09880165, + "memory(GiB)": 74.75, + "step": 25680, + "train_speed(iter/s)": 0.769245 + }, + { + "epoch": 2.39292364990689, + "grad_norm": 1.2265625, + "learning_rate": 2.1567452644305818e-05, + "loss": 0.0979923, + "memory(GiB)": 74.75, + "step": 25700, + "train_speed(iter/s)": 0.769264 + }, + { + "epoch": 2.394785847299814, + "grad_norm": 1.09375, + "learning_rate": 2.14402772906379e-05, + "loss": 0.09793384, + "memory(GiB)": 74.75, + "step": 25720, + "train_speed(iter/s)": 0.769313 + }, + { + "epoch": 2.3966480446927374, + "grad_norm": 0.8046875, + "learning_rate": 2.1313432962312284e-05, + "loss": 0.10001029, + "memory(GiB)": 74.75, + "step": 25740, + "train_speed(iter/s)": 0.769316 + }, + { + "epoch": 2.398510242085661, + "grad_norm": 1.0546875, + "learning_rate": 2.118692019381008e-05, + "loss": 0.11119788, + "memory(GiB)": 74.75, + "step": 25760, + "train_speed(iter/s)": 0.769331 + }, + { + "epoch": 2.4003724394785846, + "grad_norm": 0.8515625, + "learning_rate": 2.1060739518215333e-05, + "loss": 0.09229028, + "memory(GiB)": 74.75, + "step": 25780, + "train_speed(iter/s)": 0.769372 + }, + { + "epoch": 2.4022346368715084, + "grad_norm": 1.03125, + "learning_rate": 2.0934891467212647e-05, + "loss": 0.09729228, + "memory(GiB)": 74.75, + "step": 25800, + "train_speed(iter/s)": 0.769395 + }, + { + "epoch": 2.404096834264432, + "grad_norm": 0.9296875, + "learning_rate": 2.0809376571085238e-05, + "loss": 0.09595935, + "memory(GiB)": 74.75, + "step": 25820, + "train_speed(iter/s)": 0.769411 + }, + { + "epoch": 2.4059590316573556, + "grad_norm": 1.1484375, + "learning_rate": 2.068419535871233e-05, + "loss": 0.09269361, + "memory(GiB)": 74.75, + "step": 25840, + "train_speed(iter/s)": 0.769431 + }, + { + "epoch": 2.4078212290502794, + "grad_norm": 1.3046875, + "learning_rate": 2.055934835756722e-05, + "loss": 0.09150513, + "memory(GiB)": 74.75, + "step": 25860, + "train_speed(iter/s)": 0.769471 + }, + { + "epoch": 2.4096834264432028, + "grad_norm": 0.8515625, + "learning_rate": 2.0434836093714937e-05, + "loss": 0.09933028, + "memory(GiB)": 74.75, + "step": 25880, + "train_speed(iter/s)": 0.769502 + }, + { + "epoch": 2.4115456238361266, + "grad_norm": 1.125, + "learning_rate": 2.0310659091809948e-05, + "loss": 0.09195083, + "memory(GiB)": 74.75, + "step": 25900, + "train_speed(iter/s)": 0.76953 + }, + { + "epoch": 2.4134078212290504, + "grad_norm": 0.83984375, + "learning_rate": 2.0186817875094198e-05, + "loss": 0.10260382, + "memory(GiB)": 74.75, + "step": 25920, + "train_speed(iter/s)": 0.769561 + }, + { + "epoch": 2.4152700186219738, + "grad_norm": 1.1953125, + "learning_rate": 2.0063312965394564e-05, + "loss": 0.0909433, + "memory(GiB)": 74.75, + "step": 25940, + "train_speed(iter/s)": 0.769595 + }, + { + "epoch": 2.4171322160148976, + "grad_norm": 1.53125, + "learning_rate": 1.9940144883120958e-05, + "loss": 0.10011367, + "memory(GiB)": 74.75, + "step": 25960, + "train_speed(iter/s)": 0.769627 + }, + { + "epoch": 2.4189944134078214, + "grad_norm": 1.3203125, + "learning_rate": 1.9817314147263945e-05, + "loss": 0.10159645, + "memory(GiB)": 74.75, + "step": 25980, + "train_speed(iter/s)": 0.769625 + }, + { + "epoch": 2.4208566108007448, + "grad_norm": 1.125, + "learning_rate": 1.9694821275392682e-05, + "loss": 0.09749501, + "memory(GiB)": 74.75, + "step": 26000, + "train_speed(iter/s)": 0.769664 + }, + { + "epoch": 2.4208566108007448, + "eval_loss": 0.38522472977638245, + "eval_runtime": 76.9483, + "eval_samples_per_second": 180.446, + "eval_steps_per_second": 1.417, + "step": 26000 + }, + { + "epoch": 2.4227188081936686, + "grad_norm": 0.9609375, + "learning_rate": 1.9572666783652582e-05, + "loss": 0.09559954, + "memory(GiB)": 74.75, + "step": 26020, + "train_speed(iter/s)": 0.767242 + }, + { + "epoch": 2.4245810055865924, + "grad_norm": 1.1796875, + "learning_rate": 1.945085118676332e-05, + "loss": 0.09020438, + "memory(GiB)": 74.75, + "step": 26040, + "train_speed(iter/s)": 0.767278 + }, + { + "epoch": 2.4264432029795158, + "grad_norm": 1.0, + "learning_rate": 1.9329374998016537e-05, + "loss": 0.10400308, + "memory(GiB)": 74.75, + "step": 26060, + "train_speed(iter/s)": 0.767312 + }, + { + "epoch": 2.4283054003724396, + "grad_norm": 1.046875, + "learning_rate": 1.9208238729273732e-05, + "loss": 0.10132781, + "memory(GiB)": 74.75, + "step": 26080, + "train_speed(iter/s)": 0.767353 + }, + { + "epoch": 2.430167597765363, + "grad_norm": 1.078125, + "learning_rate": 1.9087442890964102e-05, + "loss": 0.09796514, + "memory(GiB)": 74.75, + "step": 26100, + "train_speed(iter/s)": 0.76738 + }, + { + "epoch": 2.4320297951582868, + "grad_norm": 1.015625, + "learning_rate": 1.8966987992082296e-05, + "loss": 0.0868296, + "memory(GiB)": 74.75, + "step": 26120, + "train_speed(iter/s)": 0.7674 + }, + { + "epoch": 2.4338919925512106, + "grad_norm": 1.1015625, + "learning_rate": 1.884687454018652e-05, + "loss": 0.09946807, + "memory(GiB)": 74.75, + "step": 26140, + "train_speed(iter/s)": 0.767421 + }, + { + "epoch": 2.435754189944134, + "grad_norm": 0.86328125, + "learning_rate": 1.8727103041396056e-05, + "loss": 0.08833747, + "memory(GiB)": 74.75, + "step": 26160, + "train_speed(iter/s)": 0.767423 + }, + { + "epoch": 2.4376163873370578, + "grad_norm": 0.8359375, + "learning_rate": 1.860767400038941e-05, + "loss": 0.08898092, + "memory(GiB)": 74.75, + "step": 26180, + "train_speed(iter/s)": 0.76743 + }, + { + "epoch": 2.439478584729981, + "grad_norm": 1.0078125, + "learning_rate": 1.8488587920402046e-05, + "loss": 0.08464298, + "memory(GiB)": 74.75, + "step": 26200, + "train_speed(iter/s)": 0.767463 + }, + { + "epoch": 2.441340782122905, + "grad_norm": 0.875, + "learning_rate": 1.8369845303224308e-05, + "loss": 0.09342575, + "memory(GiB)": 74.75, + "step": 26220, + "train_speed(iter/s)": 0.767475 + }, + { + "epoch": 2.4432029795158288, + "grad_norm": 0.96875, + "learning_rate": 1.8251446649199355e-05, + "loss": 0.09570937, + "memory(GiB)": 74.75, + "step": 26240, + "train_speed(iter/s)": 0.767504 + }, + { + "epoch": 2.445065176908752, + "grad_norm": 0.78515625, + "learning_rate": 1.8133392457220856e-05, + "loss": 0.08847888, + "memory(GiB)": 74.75, + "step": 26260, + "train_speed(iter/s)": 0.767526 + }, + { + "epoch": 2.446927374301676, + "grad_norm": 1.125, + "learning_rate": 1.801568322473115e-05, + "loss": 0.1068172, + "memory(GiB)": 74.75, + "step": 26280, + "train_speed(iter/s)": 0.767557 + }, + { + "epoch": 2.4487895716945998, + "grad_norm": 1.234375, + "learning_rate": 1.7898319447718993e-05, + "loss": 0.10401188, + "memory(GiB)": 74.75, + "step": 26300, + "train_speed(iter/s)": 0.767592 + }, + { + "epoch": 2.450651769087523, + "grad_norm": 1.21875, + "learning_rate": 1.778130162071753e-05, + "loss": 0.10306834, + "memory(GiB)": 74.75, + "step": 26320, + "train_speed(iter/s)": 0.767625 + }, + { + "epoch": 2.452513966480447, + "grad_norm": 0.875, + "learning_rate": 1.7664630236802073e-05, + "loss": 0.09694492, + "memory(GiB)": 74.75, + "step": 26340, + "train_speed(iter/s)": 0.767644 + }, + { + "epoch": 2.4543761638733708, + "grad_norm": 1.328125, + "learning_rate": 1.7548305787588314e-05, + "loss": 0.1060955, + "memory(GiB)": 74.75, + "step": 26360, + "train_speed(iter/s)": 0.767692 + }, + { + "epoch": 2.456238361266294, + "grad_norm": 1.109375, + "learning_rate": 1.7432328763229912e-05, + "loss": 0.09546111, + "memory(GiB)": 74.75, + "step": 26380, + "train_speed(iter/s)": 0.767709 + }, + { + "epoch": 2.458100558659218, + "grad_norm": 1.0703125, + "learning_rate": 1.731669965241668e-05, + "loss": 0.11217393, + "memory(GiB)": 74.75, + "step": 26400, + "train_speed(iter/s)": 0.767736 + }, + { + "epoch": 2.4599627560521418, + "grad_norm": 1.1171875, + "learning_rate": 1.7201418942372415e-05, + "loss": 0.09984755, + "memory(GiB)": 74.75, + "step": 26420, + "train_speed(iter/s)": 0.767754 + }, + { + "epoch": 2.461824953445065, + "grad_norm": 1.0703125, + "learning_rate": 1.708648711885281e-05, + "loss": 0.09662188, + "memory(GiB)": 74.75, + "step": 26440, + "train_speed(iter/s)": 0.767787 + }, + { + "epoch": 2.463687150837989, + "grad_norm": 0.9375, + "learning_rate": 1.69719046661436e-05, + "loss": 0.09171115, + "memory(GiB)": 74.75, + "step": 26460, + "train_speed(iter/s)": 0.767817 + }, + { + "epoch": 2.4655493482309123, + "grad_norm": 1.203125, + "learning_rate": 1.6857672067058216e-05, + "loss": 0.09821294, + "memory(GiB)": 74.75, + "step": 26480, + "train_speed(iter/s)": 0.767854 + }, + { + "epoch": 2.467411545623836, + "grad_norm": 0.734375, + "learning_rate": 1.674378980293605e-05, + "loss": 0.10482373, + "memory(GiB)": 74.75, + "step": 26500, + "train_speed(iter/s)": 0.767907 + }, + { + "epoch": 2.46927374301676, + "grad_norm": 1.078125, + "learning_rate": 1.6630258353640217e-05, + "loss": 0.08855894, + "memory(GiB)": 74.75, + "step": 26520, + "train_speed(iter/s)": 0.767921 + }, + { + "epoch": 2.4711359404096833, + "grad_norm": 0.96484375, + "learning_rate": 1.6517078197555647e-05, + "loss": 0.08362047, + "memory(GiB)": 74.75, + "step": 26540, + "train_speed(iter/s)": 0.767932 + }, + { + "epoch": 2.472998137802607, + "grad_norm": 1.0390625, + "learning_rate": 1.6404249811587056e-05, + "loss": 0.09590976, + "memory(GiB)": 74.75, + "step": 26560, + "train_speed(iter/s)": 0.767948 + }, + { + "epoch": 2.4748603351955305, + "grad_norm": 1.2265625, + "learning_rate": 1.629177367115684e-05, + "loss": 0.11099143, + "memory(GiB)": 74.75, + "step": 26580, + "train_speed(iter/s)": 0.767995 + }, + { + "epoch": 2.4767225325884543, + "grad_norm": 1.140625, + "learning_rate": 1.6179650250203216e-05, + "loss": 0.09504373, + "memory(GiB)": 74.75, + "step": 26600, + "train_speed(iter/s)": 0.768019 + }, + { + "epoch": 2.478584729981378, + "grad_norm": 0.88671875, + "learning_rate": 1.6067880021178105e-05, + "loss": 0.09536514, + "memory(GiB)": 74.75, + "step": 26620, + "train_speed(iter/s)": 0.768043 + }, + { + "epoch": 2.4804469273743015, + "grad_norm": 0.95703125, + "learning_rate": 1.5956463455045267e-05, + "loss": 0.10592208, + "memory(GiB)": 74.75, + "step": 26640, + "train_speed(iter/s)": 0.768061 + }, + { + "epoch": 2.4823091247672253, + "grad_norm": 1.015625, + "learning_rate": 1.5845401021278106e-05, + "loss": 0.10959908, + "memory(GiB)": 74.75, + "step": 26660, + "train_speed(iter/s)": 0.768072 + }, + { + "epoch": 2.484171322160149, + "grad_norm": 1.1015625, + "learning_rate": 1.5734693187857997e-05, + "loss": 0.09669304, + "memory(GiB)": 74.75, + "step": 26680, + "train_speed(iter/s)": 0.768115 + }, + { + "epoch": 2.4860335195530725, + "grad_norm": 0.96875, + "learning_rate": 1.5624340421272e-05, + "loss": 0.10210849, + "memory(GiB)": 74.75, + "step": 26700, + "train_speed(iter/s)": 0.76816 + }, + { + "epoch": 2.4878957169459963, + "grad_norm": 1.4140625, + "learning_rate": 1.5514343186511104e-05, + "loss": 0.10455359, + "memory(GiB)": 74.75, + "step": 26720, + "train_speed(iter/s)": 0.768193 + }, + { + "epoch": 2.48975791433892, + "grad_norm": 1.21875, + "learning_rate": 1.5404701947068224e-05, + "loss": 0.10247712, + "memory(GiB)": 74.75, + "step": 26740, + "train_speed(iter/s)": 0.768229 + }, + { + "epoch": 2.4916201117318435, + "grad_norm": 1.078125, + "learning_rate": 1.529541716493612e-05, + "loss": 0.09632077, + "memory(GiB)": 74.75, + "step": 26760, + "train_speed(iter/s)": 0.768262 + }, + { + "epoch": 2.4934823091247673, + "grad_norm": 1.140625, + "learning_rate": 1.518648930060571e-05, + "loss": 0.10718331, + "memory(GiB)": 74.75, + "step": 26780, + "train_speed(iter/s)": 0.768295 + }, + { + "epoch": 2.4953445065176907, + "grad_norm": 1.09375, + "learning_rate": 1.507791881306384e-05, + "loss": 0.09734278, + "memory(GiB)": 74.75, + "step": 26800, + "train_speed(iter/s)": 0.768348 + }, + { + "epoch": 2.4972067039106145, + "grad_norm": 1.1640625, + "learning_rate": 1.4969706159791564e-05, + "loss": 0.10756767, + "memory(GiB)": 74.75, + "step": 26820, + "train_speed(iter/s)": 0.768362 + }, + { + "epoch": 2.4990689013035383, + "grad_norm": 1.140625, + "learning_rate": 1.4861851796762094e-05, + "loss": 0.10674, + "memory(GiB)": 74.75, + "step": 26840, + "train_speed(iter/s)": 0.768384 + }, + { + "epoch": 2.5009310986964617, + "grad_norm": 1.0546875, + "learning_rate": 1.4754356178438954e-05, + "loss": 0.10347006, + "memory(GiB)": 74.75, + "step": 26860, + "train_speed(iter/s)": 0.768408 + }, + { + "epoch": 2.5027932960893855, + "grad_norm": 0.875, + "learning_rate": 1.4647219757774045e-05, + "loss": 0.08911729, + "memory(GiB)": 74.75, + "step": 26880, + "train_speed(iter/s)": 0.768442 + }, + { + "epoch": 2.504655493482309, + "grad_norm": 0.9765625, + "learning_rate": 1.454044298620566e-05, + "loss": 0.09411467, + "memory(GiB)": 74.75, + "step": 26900, + "train_speed(iter/s)": 0.768473 + }, + { + "epoch": 2.5065176908752327, + "grad_norm": 1.0, + "learning_rate": 1.443402631365669e-05, + "loss": 0.08597951, + "memory(GiB)": 74.75, + "step": 26920, + "train_speed(iter/s)": 0.768502 + }, + { + "epoch": 2.5083798882681565, + "grad_norm": 1.3515625, + "learning_rate": 1.4327970188532702e-05, + "loss": 0.10816082, + "memory(GiB)": 74.75, + "step": 26940, + "train_speed(iter/s)": 0.768536 + }, + { + "epoch": 2.51024208566108, + "grad_norm": 1.3046875, + "learning_rate": 1.4222275057720024e-05, + "loss": 0.08901605, + "memory(GiB)": 74.75, + "step": 26960, + "train_speed(iter/s)": 0.768546 + }, + { + "epoch": 2.5121042830540037, + "grad_norm": 1.1015625, + "learning_rate": 1.4116941366583802e-05, + "loss": 0.0930239, + "memory(GiB)": 74.75, + "step": 26980, + "train_speed(iter/s)": 0.768569 + }, + { + "epoch": 2.5139664804469275, + "grad_norm": 0.9609375, + "learning_rate": 1.4011969558966331e-05, + "loss": 0.10327775, + "memory(GiB)": 74.75, + "step": 27000, + "train_speed(iter/s)": 0.768586 + }, + { + "epoch": 2.515828677839851, + "grad_norm": 1.28125, + "learning_rate": 1.3907360077184906e-05, + "loss": 0.10326664, + "memory(GiB)": 74.75, + "step": 27020, + "train_speed(iter/s)": 0.768624 + }, + { + "epoch": 2.5176908752327747, + "grad_norm": 1.3828125, + "learning_rate": 1.3803113362030151e-05, + "loss": 0.09636109, + "memory(GiB)": 74.75, + "step": 27040, + "train_speed(iter/s)": 0.76864 + }, + { + "epoch": 2.5195530726256985, + "grad_norm": 0.99609375, + "learning_rate": 1.3699229852764172e-05, + "loss": 0.10351133, + "memory(GiB)": 74.75, + "step": 27060, + "train_speed(iter/s)": 0.768667 + }, + { + "epoch": 2.521415270018622, + "grad_norm": 1.0390625, + "learning_rate": 1.3595709987118487e-05, + "loss": 0.09553545, + "memory(GiB)": 74.75, + "step": 27080, + "train_speed(iter/s)": 0.76871 + }, + { + "epoch": 2.5232774674115457, + "grad_norm": 0.921875, + "learning_rate": 1.349255420129253e-05, + "loss": 0.08632871, + "memory(GiB)": 74.75, + "step": 27100, + "train_speed(iter/s)": 0.768728 + }, + { + "epoch": 2.5251396648044695, + "grad_norm": 0.96484375, + "learning_rate": 1.3389762929951467e-05, + "loss": 0.10044611, + "memory(GiB)": 74.75, + "step": 27120, + "train_speed(iter/s)": 0.768769 + }, + { + "epoch": 2.527001862197393, + "grad_norm": 0.8515625, + "learning_rate": 1.3287336606224576e-05, + "loss": 0.09492825, + "memory(GiB)": 74.75, + "step": 27140, + "train_speed(iter/s)": 0.76879 + }, + { + "epoch": 2.5288640595903167, + "grad_norm": 1.1640625, + "learning_rate": 1.318527566170339e-05, + "loss": 0.10425776, + "memory(GiB)": 74.75, + "step": 27160, + "train_speed(iter/s)": 0.768817 + }, + { + "epoch": 2.5307262569832405, + "grad_norm": 1.1328125, + "learning_rate": 1.3083580526439787e-05, + "loss": 0.10316639, + "memory(GiB)": 74.75, + "step": 27180, + "train_speed(iter/s)": 0.768838 + }, + { + "epoch": 2.532588454376164, + "grad_norm": 1.171875, + "learning_rate": 1.298225162894433e-05, + "loss": 0.08870911, + "memory(GiB)": 74.75, + "step": 27200, + "train_speed(iter/s)": 0.768852 + }, + { + "epoch": 2.5344506517690877, + "grad_norm": 1.0, + "learning_rate": 1.288128939618426e-05, + "loss": 0.09396772, + "memory(GiB)": 74.75, + "step": 27220, + "train_speed(iter/s)": 0.768884 + }, + { + "epoch": 2.536312849162011, + "grad_norm": 0.90625, + "learning_rate": 1.2780694253581905e-05, + "loss": 0.09058628, + "memory(GiB)": 74.75, + "step": 27240, + "train_speed(iter/s)": 0.768915 + }, + { + "epoch": 2.538175046554935, + "grad_norm": 1.171875, + "learning_rate": 1.2680466625012766e-05, + "loss": 0.09303147, + "memory(GiB)": 74.75, + "step": 27260, + "train_speed(iter/s)": 0.76895 + }, + { + "epoch": 2.5400372439478582, + "grad_norm": 1.109375, + "learning_rate": 1.2580606932803762e-05, + "loss": 0.09020774, + "memory(GiB)": 74.75, + "step": 27280, + "train_speed(iter/s)": 0.768973 + }, + { + "epoch": 2.541899441340782, + "grad_norm": 1.0546875, + "learning_rate": 1.2481115597731385e-05, + "loss": 0.0921455, + "memory(GiB)": 74.75, + "step": 27300, + "train_speed(iter/s)": 0.769017 + }, + { + "epoch": 2.543761638733706, + "grad_norm": 0.80078125, + "learning_rate": 1.2381993039020123e-05, + "loss": 0.10183475, + "memory(GiB)": 74.75, + "step": 27320, + "train_speed(iter/s)": 0.769039 + }, + { + "epoch": 2.5456238361266292, + "grad_norm": 0.828125, + "learning_rate": 1.228323967434043e-05, + "loss": 0.08973705, + "memory(GiB)": 74.75, + "step": 27340, + "train_speed(iter/s)": 0.769069 + }, + { + "epoch": 2.547486033519553, + "grad_norm": 1.0234375, + "learning_rate": 1.2184855919807148e-05, + "loss": 0.08727324, + "memory(GiB)": 74.75, + "step": 27360, + "train_speed(iter/s)": 0.769103 + }, + { + "epoch": 2.549348230912477, + "grad_norm": 1.0234375, + "learning_rate": 1.2086842189977698e-05, + "loss": 0.09381813, + "memory(GiB)": 74.75, + "step": 27380, + "train_speed(iter/s)": 0.769124 + }, + { + "epoch": 2.5512104283054002, + "grad_norm": 1.0703125, + "learning_rate": 1.1989198897850352e-05, + "loss": 0.09397247, + "memory(GiB)": 74.75, + "step": 27400, + "train_speed(iter/s)": 0.769158 + }, + { + "epoch": 2.553072625698324, + "grad_norm": 1.078125, + "learning_rate": 1.1891926454862468e-05, + "loss": 0.10348772, + "memory(GiB)": 74.75, + "step": 27420, + "train_speed(iter/s)": 0.769173 + }, + { + "epoch": 2.554934823091248, + "grad_norm": 1.0078125, + "learning_rate": 1.1795025270888726e-05, + "loss": 0.10088123, + "memory(GiB)": 74.75, + "step": 27440, + "train_speed(iter/s)": 0.769208 + }, + { + "epoch": 2.5567970204841712, + "grad_norm": 1.2890625, + "learning_rate": 1.169849575423948e-05, + "loss": 0.10310076, + "memory(GiB)": 74.75, + "step": 27460, + "train_speed(iter/s)": 0.769257 + }, + { + "epoch": 2.558659217877095, + "grad_norm": 1.0, + "learning_rate": 1.1602338311659012e-05, + "loss": 0.08614667, + "memory(GiB)": 74.75, + "step": 27480, + "train_speed(iter/s)": 0.769299 + }, + { + "epoch": 2.560521415270019, + "grad_norm": 1.203125, + "learning_rate": 1.1506553348323757e-05, + "loss": 0.10040137, + "memory(GiB)": 74.75, + "step": 27500, + "train_speed(iter/s)": 0.769316 + }, + { + "epoch": 2.5623836126629422, + "grad_norm": 1.0234375, + "learning_rate": 1.1411141267840675e-05, + "loss": 0.08545375, + "memory(GiB)": 74.75, + "step": 27520, + "train_speed(iter/s)": 0.769353 + }, + { + "epoch": 2.564245810055866, + "grad_norm": 0.95703125, + "learning_rate": 1.131610247224555e-05, + "loss": 0.1016354, + "memory(GiB)": 74.75, + "step": 27540, + "train_speed(iter/s)": 0.769373 + }, + { + "epoch": 2.5661080074487894, + "grad_norm": 0.9609375, + "learning_rate": 1.1221437362001153e-05, + "loss": 0.09240752, + "memory(GiB)": 74.75, + "step": 27560, + "train_speed(iter/s)": 0.769391 + }, + { + "epoch": 2.5679702048417132, + "grad_norm": 1.0625, + "learning_rate": 1.1127146335995774e-05, + "loss": 0.09214135, + "memory(GiB)": 74.75, + "step": 27580, + "train_speed(iter/s)": 0.76942 + }, + { + "epoch": 2.5698324022346366, + "grad_norm": 1.1015625, + "learning_rate": 1.1033229791541422e-05, + "loss": 0.10229049, + "memory(GiB)": 74.75, + "step": 27600, + "train_speed(iter/s)": 0.769439 + }, + { + "epoch": 2.5716945996275604, + "grad_norm": 1.4609375, + "learning_rate": 1.093968812437206e-05, + "loss": 0.11024711, + "memory(GiB)": 74.75, + "step": 27620, + "train_speed(iter/s)": 0.769455 + }, + { + "epoch": 2.5735567970204842, + "grad_norm": 1.09375, + "learning_rate": 1.0846521728642201e-05, + "loss": 0.10305281, + "memory(GiB)": 74.75, + "step": 27640, + "train_speed(iter/s)": 0.769477 + }, + { + "epoch": 2.5754189944134076, + "grad_norm": 0.95703125, + "learning_rate": 1.0753730996924926e-05, + "loss": 0.10268775, + "memory(GiB)": 74.75, + "step": 27660, + "train_speed(iter/s)": 0.769492 + }, + { + "epoch": 2.5772811918063314, + "grad_norm": 0.83203125, + "learning_rate": 1.0661316320210524e-05, + "loss": 0.08912667, + "memory(GiB)": 74.75, + "step": 27680, + "train_speed(iter/s)": 0.769514 + }, + { + "epoch": 2.5791433891992552, + "grad_norm": 1.03125, + "learning_rate": 1.0569278087904588e-05, + "loss": 0.09018875, + "memory(GiB)": 74.75, + "step": 27700, + "train_speed(iter/s)": 0.769546 + }, + { + "epoch": 2.5810055865921786, + "grad_norm": 1.15625, + "learning_rate": 1.0477616687826596e-05, + "loss": 0.08439552, + "memory(GiB)": 74.75, + "step": 27720, + "train_speed(iter/s)": 0.76954 + }, + { + "epoch": 2.5828677839851024, + "grad_norm": 0.84765625, + "learning_rate": 1.038633250620813e-05, + "loss": 0.10151093, + "memory(GiB)": 74.75, + "step": 27740, + "train_speed(iter/s)": 0.76954 + }, + { + "epoch": 2.5847299813780262, + "grad_norm": 1.0546875, + "learning_rate": 1.029542592769126e-05, + "loss": 0.10534508, + "memory(GiB)": 74.75, + "step": 27760, + "train_speed(iter/s)": 0.769567 + }, + { + "epoch": 2.5865921787709496, + "grad_norm": 1.03125, + "learning_rate": 1.0204897335327047e-05, + "loss": 0.09673885, + "memory(GiB)": 74.75, + "step": 27780, + "train_speed(iter/s)": 0.769612 + }, + { + "epoch": 2.5884543761638734, + "grad_norm": 1.453125, + "learning_rate": 1.0114747110573753e-05, + "loss": 0.10991576, + "memory(GiB)": 74.75, + "step": 27800, + "train_speed(iter/s)": 0.769639 + }, + { + "epoch": 2.5903165735567972, + "grad_norm": 1.1796875, + "learning_rate": 1.0024975633295386e-05, + "loss": 0.08816065, + "memory(GiB)": 74.75, + "step": 27820, + "train_speed(iter/s)": 0.769669 + }, + { + "epoch": 2.5921787709497206, + "grad_norm": 0.87890625, + "learning_rate": 9.935583281759986e-06, + "loss": 0.10151916, + "memory(GiB)": 74.75, + "step": 27840, + "train_speed(iter/s)": 0.769699 + }, + { + "epoch": 2.5940409683426444, + "grad_norm": 0.96484375, + "learning_rate": 9.846570432638147e-06, + "loss": 0.10038137, + "memory(GiB)": 74.75, + "step": 27860, + "train_speed(iter/s)": 0.769711 + }, + { + "epoch": 2.5959031657355682, + "grad_norm": 0.98828125, + "learning_rate": 9.757937461001244e-06, + "loss": 0.08528215, + "memory(GiB)": 74.75, + "step": 27880, + "train_speed(iter/s)": 0.769713 + }, + { + "epoch": 2.5977653631284916, + "grad_norm": 0.85546875, + "learning_rate": 9.669684740320096e-06, + "loss": 0.08982363, + "memory(GiB)": 74.75, + "step": 27900, + "train_speed(iter/s)": 0.76973 + }, + { + "epoch": 2.5996275605214154, + "grad_norm": 1.125, + "learning_rate": 9.581812642463206e-06, + "loss": 0.11246864, + "memory(GiB)": 74.75, + "step": 27920, + "train_speed(iter/s)": 0.769746 + }, + { + "epoch": 2.601489757914339, + "grad_norm": 1.0859375, + "learning_rate": 9.494321537695216e-06, + "loss": 0.100134, + "memory(GiB)": 74.75, + "step": 27940, + "train_speed(iter/s)": 0.769769 + }, + { + "epoch": 2.6033519553072626, + "grad_norm": 1.0546875, + "learning_rate": 9.407211794675508e-06, + "loss": 0.09557046, + "memory(GiB)": 74.75, + "step": 27960, + "train_speed(iter/s)": 0.76978 + }, + { + "epoch": 2.605214152700186, + "grad_norm": 0.83984375, + "learning_rate": 9.32048378045638e-06, + "loss": 0.09010305, + "memory(GiB)": 74.75, + "step": 27980, + "train_speed(iter/s)": 0.769779 + }, + { + "epoch": 2.60707635009311, + "grad_norm": 0.91015625, + "learning_rate": 9.234137860481796e-06, + "loss": 0.09046291, + "memory(GiB)": 45.75, + "step": 28000, + "train_speed(iter/s)": 0.769799 + }, + { + "epoch": 2.60707635009311, + "eval_loss": 0.38696086406707764, + "eval_runtime": 77.0234, + "eval_samples_per_second": 180.27, + "eval_steps_per_second": 1.415, + "step": 28000 + }, + { + "epoch": 2.6089385474860336, + "grad_norm": 0.97265625, + "learning_rate": 9.148174398585564e-06, + "loss": 0.09225785, + "memory(GiB)": 45.75, + "step": 28020, + "train_speed(iter/s)": 0.767502 + }, + { + "epoch": 2.610800744878957, + "grad_norm": 0.89453125, + "learning_rate": 9.062593756990012e-06, + "loss": 0.10355859, + "memory(GiB)": 45.75, + "step": 28040, + "train_speed(iter/s)": 0.767513 + }, + { + "epoch": 2.612662942271881, + "grad_norm": 1.0546875, + "learning_rate": 8.977396296304385e-06, + "loss": 0.09487821, + "memory(GiB)": 45.75, + "step": 28060, + "train_speed(iter/s)": 0.767532 + }, + { + "epoch": 2.6145251396648046, + "grad_norm": 0.86328125, + "learning_rate": 8.892582375523295e-06, + "loss": 0.08763714, + "memory(GiB)": 45.75, + "step": 28080, + "train_speed(iter/s)": 0.767557 + }, + { + "epoch": 2.616387337057728, + "grad_norm": 0.78515625, + "learning_rate": 8.808152352025267e-06, + "loss": 0.09426429, + "memory(GiB)": 45.75, + "step": 28100, + "train_speed(iter/s)": 0.767589 + }, + { + "epoch": 2.618249534450652, + "grad_norm": 1.0078125, + "learning_rate": 8.7241065815712e-06, + "loss": 0.10946054, + "memory(GiB)": 45.75, + "step": 28120, + "train_speed(iter/s)": 0.767604 + }, + { + "epoch": 2.6201117318435756, + "grad_norm": 1.03125, + "learning_rate": 8.640445418302856e-06, + "loss": 0.1012117, + "memory(GiB)": 45.75, + "step": 28140, + "train_speed(iter/s)": 0.767624 + }, + { + "epoch": 2.621973929236499, + "grad_norm": 1.1640625, + "learning_rate": 8.557169214741435e-06, + "loss": 0.09337143, + "memory(GiB)": 45.75, + "step": 28160, + "train_speed(iter/s)": 0.767651 + }, + { + "epoch": 2.623836126629423, + "grad_norm": 1.484375, + "learning_rate": 8.474278321785988e-06, + "loss": 0.08732111, + "memory(GiB)": 45.75, + "step": 28180, + "train_speed(iter/s)": 0.767693 + }, + { + "epoch": 2.6256983240223466, + "grad_norm": 1.09375, + "learning_rate": 8.391773088712018e-06, + "loss": 0.09296905, + "memory(GiB)": 45.75, + "step": 28200, + "train_speed(iter/s)": 0.767704 + }, + { + "epoch": 2.62756052141527, + "grad_norm": 0.98046875, + "learning_rate": 8.309653863169964e-06, + "loss": 0.09660787, + "memory(GiB)": 45.75, + "step": 28220, + "train_speed(iter/s)": 0.767745 + }, + { + "epoch": 2.629422718808194, + "grad_norm": 1.0078125, + "learning_rate": 8.22792099118379e-06, + "loss": 0.08955661, + "memory(GiB)": 45.75, + "step": 28240, + "train_speed(iter/s)": 0.767763 + }, + { + "epoch": 2.631284916201117, + "grad_norm": 0.890625, + "learning_rate": 8.146574817149411e-06, + "loss": 0.09369433, + "memory(GiB)": 45.75, + "step": 28260, + "train_speed(iter/s)": 0.767786 + }, + { + "epoch": 2.633147113594041, + "grad_norm": 0.98046875, + "learning_rate": 8.065615683833462e-06, + "loss": 0.09985462, + "memory(GiB)": 45.75, + "step": 28280, + "train_speed(iter/s)": 0.767812 + }, + { + "epoch": 2.635009310986965, + "grad_norm": 1.078125, + "learning_rate": 7.985043932371539e-06, + "loss": 0.09353604, + "memory(GiB)": 45.75, + "step": 28300, + "train_speed(iter/s)": 0.767841 + }, + { + "epoch": 2.636871508379888, + "grad_norm": 1.0078125, + "learning_rate": 7.904859902267114e-06, + "loss": 0.08616074, + "memory(GiB)": 45.75, + "step": 28320, + "train_speed(iter/s)": 0.767875 + }, + { + "epoch": 2.638733705772812, + "grad_norm": 1.234375, + "learning_rate": 7.825063931389764e-06, + "loss": 0.09571852, + "memory(GiB)": 45.75, + "step": 28340, + "train_speed(iter/s)": 0.767919 + }, + { + "epoch": 2.6405959031657353, + "grad_norm": 1.328125, + "learning_rate": 7.745656355974018e-06, + "loss": 0.10770638, + "memory(GiB)": 45.75, + "step": 28360, + "train_speed(iter/s)": 0.767942 + }, + { + "epoch": 2.642458100558659, + "grad_norm": 1.1015625, + "learning_rate": 7.666637510617791e-06, + "loss": 0.08916503, + "memory(GiB)": 45.75, + "step": 28380, + "train_speed(iter/s)": 0.767976 + }, + { + "epoch": 2.644320297951583, + "grad_norm": 1.15625, + "learning_rate": 7.58800772828101e-06, + "loss": 0.1024822, + "memory(GiB)": 45.75, + "step": 28400, + "train_speed(iter/s)": 0.768012 + }, + { + "epoch": 2.6461824953445063, + "grad_norm": 1.109375, + "learning_rate": 7.509767340284235e-06, + "loss": 0.08160416, + "memory(GiB)": 45.75, + "step": 28420, + "train_speed(iter/s)": 0.768045 + }, + { + "epoch": 2.64804469273743, + "grad_norm": 1.046875, + "learning_rate": 7.431916676307238e-06, + "loss": 0.09086616, + "memory(GiB)": 45.75, + "step": 28440, + "train_speed(iter/s)": 0.768057 + }, + { + "epoch": 2.649906890130354, + "grad_norm": 1.0859375, + "learning_rate": 7.354456064387605e-06, + "loss": 0.10261321, + "memory(GiB)": 45.75, + "step": 28460, + "train_speed(iter/s)": 0.768083 + }, + { + "epoch": 2.6517690875232773, + "grad_norm": 1.2734375, + "learning_rate": 7.277385830919381e-06, + "loss": 0.09744512, + "memory(GiB)": 45.75, + "step": 28480, + "train_speed(iter/s)": 0.768101 + }, + { + "epoch": 2.653631284916201, + "grad_norm": 0.98046875, + "learning_rate": 7.200706300651705e-06, + "loss": 0.09409308, + "memory(GiB)": 45.75, + "step": 28500, + "train_speed(iter/s)": 0.768142 + }, + { + "epoch": 2.655493482309125, + "grad_norm": 1.1484375, + "learning_rate": 7.124417796687366e-06, + "loss": 0.09326741, + "memory(GiB)": 45.75, + "step": 28520, + "train_speed(iter/s)": 0.768152 + }, + { + "epoch": 2.6573556797020483, + "grad_norm": 0.78125, + "learning_rate": 7.048520640481526e-06, + "loss": 0.097636, + "memory(GiB)": 45.75, + "step": 28540, + "train_speed(iter/s)": 0.768179 + }, + { + "epoch": 2.659217877094972, + "grad_norm": 1.125, + "learning_rate": 6.973015151840334e-06, + "loss": 0.09104769, + "memory(GiB)": 45.75, + "step": 28560, + "train_speed(iter/s)": 0.768207 + }, + { + "epoch": 2.661080074487896, + "grad_norm": 1.125, + "learning_rate": 6.897901648919591e-06, + "loss": 0.1032981, + "memory(GiB)": 45.75, + "step": 28580, + "train_speed(iter/s)": 0.768224 + }, + { + "epoch": 2.6629422718808193, + "grad_norm": 0.8828125, + "learning_rate": 6.8231804482233655e-06, + "loss": 0.08662249, + "memory(GiB)": 45.75, + "step": 28600, + "train_speed(iter/s)": 0.768234 + }, + { + "epoch": 2.664804469273743, + "grad_norm": 1.1640625, + "learning_rate": 6.74885186460269e-06, + "loss": 0.1023387, + "memory(GiB)": 45.75, + "step": 28620, + "train_speed(iter/s)": 0.768249 + }, + { + "epoch": 2.6666666666666665, + "grad_norm": 1.2890625, + "learning_rate": 6.674916211254289e-06, + "loss": 0.09208263, + "memory(GiB)": 45.75, + "step": 28640, + "train_speed(iter/s)": 0.768252 + }, + { + "epoch": 2.6685288640595903, + "grad_norm": 1.1875, + "learning_rate": 6.601373799719135e-06, + "loss": 0.09997191, + "memory(GiB)": 45.75, + "step": 28660, + "train_speed(iter/s)": 0.768271 + }, + { + "epoch": 2.6703910614525137, + "grad_norm": 1.0625, + "learning_rate": 6.528224939881256e-06, + "loss": 0.09428074, + "memory(GiB)": 45.75, + "step": 28680, + "train_speed(iter/s)": 0.768303 + }, + { + "epoch": 2.6722532588454375, + "grad_norm": 0.984375, + "learning_rate": 6.455469939966352e-06, + "loss": 0.09120364, + "memory(GiB)": 45.75, + "step": 28700, + "train_speed(iter/s)": 0.768333 + }, + { + "epoch": 2.6741154562383613, + "grad_norm": 1.1171875, + "learning_rate": 6.383109106540552e-06, + "loss": 0.09465919, + "memory(GiB)": 45.75, + "step": 28720, + "train_speed(iter/s)": 0.768359 + }, + { + "epoch": 2.6759776536312847, + "grad_norm": 1.515625, + "learning_rate": 6.311142744509058e-06, + "loss": 0.10171504, + "memory(GiB)": 45.75, + "step": 28740, + "train_speed(iter/s)": 0.768388 + }, + { + "epoch": 2.6778398510242085, + "grad_norm": 0.9921875, + "learning_rate": 6.239571157114921e-06, + "loss": 0.08548477, + "memory(GiB)": 45.75, + "step": 28760, + "train_speed(iter/s)": 0.768436 + }, + { + "epoch": 2.6797020484171323, + "grad_norm": 1.140625, + "learning_rate": 6.168394645937725e-06, + "loss": 0.10516393, + "memory(GiB)": 45.75, + "step": 28780, + "train_speed(iter/s)": 0.768465 + }, + { + "epoch": 2.6815642458100557, + "grad_norm": 0.96484375, + "learning_rate": 6.0976135108923636e-06, + "loss": 0.10106995, + "memory(GiB)": 45.75, + "step": 28800, + "train_speed(iter/s)": 0.768484 + }, + { + "epoch": 2.6834264432029795, + "grad_norm": 0.83984375, + "learning_rate": 6.027228050227718e-06, + "loss": 0.08965033, + "memory(GiB)": 45.75, + "step": 28820, + "train_speed(iter/s)": 0.768529 + }, + { + "epoch": 2.6852886405959033, + "grad_norm": 0.8828125, + "learning_rate": 5.957238560525391e-06, + "loss": 0.09630392, + "memory(GiB)": 45.75, + "step": 28840, + "train_speed(iter/s)": 0.768562 + }, + { + "epoch": 2.6871508379888267, + "grad_norm": 1.15625, + "learning_rate": 5.887645336698566e-06, + "loss": 0.08954536, + "memory(GiB)": 45.75, + "step": 28860, + "train_speed(iter/s)": 0.768579 + }, + { + "epoch": 2.6890130353817505, + "grad_norm": 0.85546875, + "learning_rate": 5.818448671990628e-06, + "loss": 0.10974171, + "memory(GiB)": 45.75, + "step": 28880, + "train_speed(iter/s)": 0.768584 + }, + { + "epoch": 2.6908752327746743, + "grad_norm": 1.15625, + "learning_rate": 5.7496488579740105e-06, + "loss": 0.10110396, + "memory(GiB)": 45.75, + "step": 28900, + "train_speed(iter/s)": 0.768602 + }, + { + "epoch": 2.6927374301675977, + "grad_norm": 0.98046875, + "learning_rate": 5.681246184548972e-06, + "loss": 0.11610439, + "memory(GiB)": 45.75, + "step": 28920, + "train_speed(iter/s)": 0.768609 + }, + { + "epoch": 2.6945996275605215, + "grad_norm": 0.9609375, + "learning_rate": 5.613240939942277e-06, + "loss": 0.09422116, + "memory(GiB)": 45.75, + "step": 28940, + "train_speed(iter/s)": 0.768625 + }, + { + "epoch": 2.6964618249534453, + "grad_norm": 0.75, + "learning_rate": 5.5456334107061635e-06, + "loss": 0.10199471, + "memory(GiB)": 45.75, + "step": 28960, + "train_speed(iter/s)": 0.768654 + }, + { + "epoch": 2.6983240223463687, + "grad_norm": 1.1328125, + "learning_rate": 5.478423881716899e-06, + "loss": 0.0880377, + "memory(GiB)": 45.75, + "step": 28980, + "train_speed(iter/s)": 0.768691 + }, + { + "epoch": 2.7001862197392925, + "grad_norm": 1.09375, + "learning_rate": 5.4116126361738125e-06, + "loss": 0.08939071, + "memory(GiB)": 45.75, + "step": 29000, + "train_speed(iter/s)": 0.76867 + }, + { + "epoch": 2.702048417132216, + "grad_norm": 1.1484375, + "learning_rate": 5.345199955597924e-06, + "loss": 0.10153615, + "memory(GiB)": 45.75, + "step": 29020, + "train_speed(iter/s)": 0.768684 + }, + { + "epoch": 2.7039106145251397, + "grad_norm": 0.8359375, + "learning_rate": 5.279186119830892e-06, + "loss": 0.09276768, + "memory(GiB)": 45.75, + "step": 29040, + "train_speed(iter/s)": 0.768714 + }, + { + "epoch": 2.705772811918063, + "grad_norm": 0.890625, + "learning_rate": 5.21357140703369e-06, + "loss": 0.09366176, + "memory(GiB)": 45.75, + "step": 29060, + "train_speed(iter/s)": 0.768739 + }, + { + "epoch": 2.707635009310987, + "grad_norm": 1.125, + "learning_rate": 5.148356093685591e-06, + "loss": 0.09835015, + "memory(GiB)": 45.75, + "step": 29080, + "train_speed(iter/s)": 0.768747 + }, + { + "epoch": 2.7094972067039107, + "grad_norm": 1.0078125, + "learning_rate": 5.083540454582891e-06, + "loss": 0.09732973, + "memory(GiB)": 45.75, + "step": 29100, + "train_speed(iter/s)": 0.768793 + }, + { + "epoch": 2.711359404096834, + "grad_norm": 0.90625, + "learning_rate": 5.019124762837801e-06, + "loss": 0.09556629, + "memory(GiB)": 45.75, + "step": 29120, + "train_speed(iter/s)": 0.768826 + }, + { + "epoch": 2.713221601489758, + "grad_norm": 0.84375, + "learning_rate": 4.95510928987728e-06, + "loss": 0.10411631, + "memory(GiB)": 45.75, + "step": 29140, + "train_speed(iter/s)": 0.768842 + }, + { + "epoch": 2.7150837988826817, + "grad_norm": 1.4765625, + "learning_rate": 4.891494305441868e-06, + "loss": 0.09335988, + "memory(GiB)": 45.75, + "step": 29160, + "train_speed(iter/s)": 0.768872 + }, + { + "epoch": 2.716945996275605, + "grad_norm": 0.953125, + "learning_rate": 4.8282800775846345e-06, + "loss": 0.08496674, + "memory(GiB)": 45.75, + "step": 29180, + "train_speed(iter/s)": 0.768898 + }, + { + "epoch": 2.718808193668529, + "grad_norm": 0.984375, + "learning_rate": 4.765466872669921e-06, + "loss": 0.10083444, + "memory(GiB)": 45.75, + "step": 29200, + "train_speed(iter/s)": 0.768937 + }, + { + "epoch": 2.7206703910614527, + "grad_norm": 1.3046875, + "learning_rate": 4.703054955372344e-06, + "loss": 0.11718053, + "memory(GiB)": 45.75, + "step": 29220, + "train_speed(iter/s)": 0.768969 + }, + { + "epoch": 2.722532588454376, + "grad_norm": 0.9453125, + "learning_rate": 4.6410445886755914e-06, + "loss": 0.09782855, + "memory(GiB)": 45.75, + "step": 29240, + "train_speed(iter/s)": 0.768995 + }, + { + "epoch": 2.7243947858473, + "grad_norm": 0.73828125, + "learning_rate": 4.579436033871342e-06, + "loss": 0.09005135, + "memory(GiB)": 45.75, + "step": 29260, + "train_speed(iter/s)": 0.76898 + }, + { + "epoch": 2.7262569832402237, + "grad_norm": 1.125, + "learning_rate": 4.518229550558218e-06, + "loss": 0.1048403, + "memory(GiB)": 45.75, + "step": 29280, + "train_speed(iter/s)": 0.769008 + }, + { + "epoch": 2.728119180633147, + "grad_norm": 0.90234375, + "learning_rate": 4.457425396640558e-06, + "loss": 0.08753203, + "memory(GiB)": 45.75, + "step": 29300, + "train_speed(iter/s)": 0.769034 + }, + { + "epoch": 2.729981378026071, + "grad_norm": 1.140625, + "learning_rate": 4.397023828327496e-06, + "loss": 0.1012768, + "memory(GiB)": 45.75, + "step": 29320, + "train_speed(iter/s)": 0.76906 + }, + { + "epoch": 2.7318435754189943, + "grad_norm": 1.15625, + "learning_rate": 4.337025100131764e-06, + "loss": 0.09988154, + "memory(GiB)": 45.75, + "step": 29340, + "train_speed(iter/s)": 0.769086 + }, + { + "epoch": 2.733705772811918, + "grad_norm": 0.74609375, + "learning_rate": 4.277429464868654e-06, + "loss": 0.10575411, + "memory(GiB)": 45.75, + "step": 29360, + "train_speed(iter/s)": 0.769107 + }, + { + "epoch": 2.7355679702048414, + "grad_norm": 1.0546875, + "learning_rate": 4.2182371736549375e-06, + "loss": 0.0899995, + "memory(GiB)": 45.75, + "step": 29380, + "train_speed(iter/s)": 0.769123 + }, + { + "epoch": 2.7374301675977653, + "grad_norm": 0.9375, + "learning_rate": 4.159448475907879e-06, + "loss": 0.08668512, + "memory(GiB)": 45.75, + "step": 29400, + "train_speed(iter/s)": 0.76917 + }, + { + "epoch": 2.739292364990689, + "grad_norm": 0.98046875, + "learning_rate": 4.10106361934407e-06, + "loss": 0.09873737, + "memory(GiB)": 45.75, + "step": 29420, + "train_speed(iter/s)": 0.769201 + }, + { + "epoch": 2.7411545623836124, + "grad_norm": 1.1953125, + "learning_rate": 4.043082849978475e-06, + "loss": 0.11074016, + "memory(GiB)": 45.75, + "step": 29440, + "train_speed(iter/s)": 0.769209 + }, + { + "epoch": 2.7430167597765363, + "grad_norm": 1.2265625, + "learning_rate": 3.985506412123397e-06, + "loss": 0.10686636, + "memory(GiB)": 45.75, + "step": 29460, + "train_speed(iter/s)": 0.769252 + }, + { + "epoch": 2.74487895716946, + "grad_norm": 0.8203125, + "learning_rate": 3.928334548387313e-06, + "loss": 0.09317576, + "memory(GiB)": 45.75, + "step": 29480, + "train_speed(iter/s)": 0.769275 + }, + { + "epoch": 2.7467411545623834, + "grad_norm": 1.3671875, + "learning_rate": 3.8715674996740894e-06, + "loss": 0.10190638, + "memory(GiB)": 45.75, + "step": 29500, + "train_speed(iter/s)": 0.769309 + }, + { + "epoch": 2.7486033519553073, + "grad_norm": 1.109375, + "learning_rate": 3.8152055051817405e-06, + "loss": 0.1021665, + "memory(GiB)": 45.75, + "step": 29520, + "train_speed(iter/s)": 0.76935 + }, + { + "epoch": 2.750465549348231, + "grad_norm": 1.875, + "learning_rate": 3.7592488024015272e-06, + "loss": 0.09051661, + "memory(GiB)": 45.75, + "step": 29540, + "train_speed(iter/s)": 0.769382 + }, + { + "epoch": 2.7523277467411544, + "grad_norm": 0.9921875, + "learning_rate": 3.7036976271169864e-06, + "loss": 0.10837436, + "memory(GiB)": 45.75, + "step": 29560, + "train_speed(iter/s)": 0.769402 + }, + { + "epoch": 2.7541899441340782, + "grad_norm": 1.125, + "learning_rate": 3.6485522134028005e-06, + "loss": 0.09573312, + "memory(GiB)": 45.75, + "step": 29580, + "train_speed(iter/s)": 0.769426 + }, + { + "epoch": 2.756052141527002, + "grad_norm": 1.2265625, + "learning_rate": 3.593812793624041e-06, + "loss": 0.10327983, + "memory(GiB)": 45.75, + "step": 29600, + "train_speed(iter/s)": 0.76945 + }, + { + "epoch": 2.7579143389199254, + "grad_norm": 1.1015625, + "learning_rate": 3.5394795984349138e-06, + "loss": 0.09151828, + "memory(GiB)": 45.75, + "step": 29620, + "train_speed(iter/s)": 0.76947 + }, + { + "epoch": 2.7597765363128492, + "grad_norm": 0.8203125, + "learning_rate": 3.485552856778007e-06, + "loss": 0.09154891, + "memory(GiB)": 45.75, + "step": 29640, + "train_speed(iter/s)": 0.769473 + }, + { + "epoch": 2.761638733705773, + "grad_norm": 1.0, + "learning_rate": 3.432032795883222e-06, + "loss": 0.086561, + "memory(GiB)": 45.75, + "step": 29660, + "train_speed(iter/s)": 0.769506 + }, + { + "epoch": 2.7635009310986964, + "grad_norm": 1.296875, + "learning_rate": 3.3789196412668424e-06, + "loss": 0.08920512, + "memory(GiB)": 45.75, + "step": 29680, + "train_speed(iter/s)": 0.769517 + }, + { + "epoch": 2.7653631284916202, + "grad_norm": 1.0078125, + "learning_rate": 3.326213616730578e-06, + "loss": 0.09615018, + "memory(GiB)": 45.75, + "step": 29700, + "train_speed(iter/s)": 0.769545 + }, + { + "epoch": 2.7672253258845436, + "grad_norm": 0.93359375, + "learning_rate": 3.273914944360601e-06, + "loss": 0.08047535, + "memory(GiB)": 45.75, + "step": 29720, + "train_speed(iter/s)": 0.769564 + }, + { + "epoch": 2.7690875232774674, + "grad_norm": 1.109375, + "learning_rate": 3.2220238445266892e-06, + "loss": 0.08651033, + "memory(GiB)": 45.75, + "step": 29740, + "train_speed(iter/s)": 0.769571 + }, + { + "epoch": 2.770949720670391, + "grad_norm": 1.1484375, + "learning_rate": 3.1705405358811723e-06, + "loss": 0.09964473, + "memory(GiB)": 45.75, + "step": 29760, + "train_speed(iter/s)": 0.76961 + }, + { + "epoch": 2.7728119180633146, + "grad_norm": 1.140625, + "learning_rate": 3.1194652353581545e-06, + "loss": 0.09600849, + "memory(GiB)": 45.75, + "step": 29780, + "train_speed(iter/s)": 0.769634 + }, + { + "epoch": 2.7746741154562384, + "grad_norm": 1.1015625, + "learning_rate": 3.068798158172437e-06, + "loss": 0.10160735, + "memory(GiB)": 45.75, + "step": 29800, + "train_speed(iter/s)": 0.769656 + }, + { + "epoch": 2.776536312849162, + "grad_norm": 0.8828125, + "learning_rate": 3.0185395178187857e-06, + "loss": 0.10089248, + "memory(GiB)": 45.75, + "step": 29820, + "train_speed(iter/s)": 0.769644 + }, + { + "epoch": 2.7783985102420856, + "grad_norm": 1.0234375, + "learning_rate": 2.968689526070889e-06, + "loss": 0.09260511, + "memory(GiB)": 45.75, + "step": 29840, + "train_speed(iter/s)": 0.769688 + }, + { + "epoch": 2.7802607076350094, + "grad_norm": 1.0546875, + "learning_rate": 2.9192483929805335e-06, + "loss": 0.09912217, + "memory(GiB)": 45.75, + "step": 29860, + "train_speed(iter/s)": 0.769712 + }, + { + "epoch": 2.782122905027933, + "grad_norm": 0.83984375, + "learning_rate": 2.870216326876729e-06, + "loss": 0.09143957, + "memory(GiB)": 45.75, + "step": 29880, + "train_speed(iter/s)": 0.769748 + }, + { + "epoch": 2.7839851024208566, + "grad_norm": 1.0, + "learning_rate": 2.821593534364786e-06, + "loss": 0.10889904, + "memory(GiB)": 45.75, + "step": 29900, + "train_speed(iter/s)": 0.769769 + }, + { + "epoch": 2.7858472998137804, + "grad_norm": 0.9921875, + "learning_rate": 2.7733802203254832e-06, + "loss": 0.10050044, + "memory(GiB)": 45.75, + "step": 29920, + "train_speed(iter/s)": 0.769786 + }, + { + "epoch": 2.787709497206704, + "grad_norm": 1.328125, + "learning_rate": 2.7255765879141693e-06, + "loss": 0.08751726, + "memory(GiB)": 45.75, + "step": 29940, + "train_speed(iter/s)": 0.769797 + }, + { + "epoch": 2.7895716945996276, + "grad_norm": 1.1015625, + "learning_rate": 2.6781828385599394e-06, + "loss": 0.09062733, + "memory(GiB)": 45.75, + "step": 29960, + "train_speed(iter/s)": 0.769807 + }, + { + "epoch": 2.7914338919925514, + "grad_norm": 0.9375, + "learning_rate": 2.631199171964771e-06, + "loss": 0.11372201, + "memory(GiB)": 45.75, + "step": 29980, + "train_speed(iter/s)": 0.769828 + }, + { + "epoch": 2.793296089385475, + "grad_norm": 1.0625, + "learning_rate": 2.5846257861027014e-06, + "loss": 0.09804424, + "memory(GiB)": 45.75, + "step": 30000, + "train_speed(iter/s)": 0.769849 + }, + { + "epoch": 2.793296089385475, + "eval_loss": 0.3881955146789551, + "eval_runtime": 77.0253, + "eval_samples_per_second": 180.266, + "eval_steps_per_second": 1.415, + "step": 30000 + }, + { + "epoch": 2.7951582867783986, + "grad_norm": 1.2421875, + "learning_rate": 2.5384628772189388e-06, + "loss": 0.08849491, + "memory(GiB)": 45.75, + "step": 30020, + "train_speed(iter/s)": 0.767704 + }, + { + "epoch": 2.7970204841713224, + "grad_norm": 1.2421875, + "learning_rate": 2.492710639829121e-06, + "loss": 0.10444663, + "memory(GiB)": 45.75, + "step": 30040, + "train_speed(iter/s)": 0.767745 + }, + { + "epoch": 2.798882681564246, + "grad_norm": 1.03125, + "learning_rate": 2.4473692667184133e-06, + "loss": 0.09703537, + "memory(GiB)": 45.75, + "step": 30060, + "train_speed(iter/s)": 0.767778 + }, + { + "epoch": 2.8007448789571696, + "grad_norm": 0.828125, + "learning_rate": 2.402438948940744e-06, + "loss": 0.09352793, + "memory(GiB)": 45.75, + "step": 30080, + "train_speed(iter/s)": 0.767825 + }, + { + "epoch": 2.802607076350093, + "grad_norm": 0.97265625, + "learning_rate": 2.3579198758179934e-06, + "loss": 0.10403392, + "memory(GiB)": 45.75, + "step": 30100, + "train_speed(iter/s)": 0.767855 + }, + { + "epoch": 2.804469273743017, + "grad_norm": 1.109375, + "learning_rate": 2.3138122349391612e-06, + "loss": 0.08864436, + "memory(GiB)": 45.75, + "step": 30120, + "train_speed(iter/s)": 0.767874 + }, + { + "epoch": 2.80633147113594, + "grad_norm": 0.890625, + "learning_rate": 2.2701162121596454e-06, + "loss": 0.10232602, + "memory(GiB)": 45.75, + "step": 30140, + "train_speed(iter/s)": 0.767912 + }, + { + "epoch": 2.808193668528864, + "grad_norm": 1.125, + "learning_rate": 2.226831991600398e-06, + "loss": 0.09759098, + "memory(GiB)": 45.75, + "step": 30160, + "train_speed(iter/s)": 0.767926 + }, + { + "epoch": 2.810055865921788, + "grad_norm": 1.203125, + "learning_rate": 2.183959755647147e-06, + "loss": 0.08362782, + "memory(GiB)": 45.75, + "step": 30180, + "train_speed(iter/s)": 0.767937 + }, + { + "epoch": 2.811918063314711, + "grad_norm": 1.140625, + "learning_rate": 2.141499684949688e-06, + "loss": 0.09115112, + "memory(GiB)": 45.75, + "step": 30200, + "train_speed(iter/s)": 0.767948 + }, + { + "epoch": 2.813780260707635, + "grad_norm": 0.8515625, + "learning_rate": 2.0994519584210728e-06, + "loss": 0.09945311, + "memory(GiB)": 45.75, + "step": 30220, + "train_speed(iter/s)": 0.767951 + }, + { + "epoch": 2.815642458100559, + "grad_norm": 0.98046875, + "learning_rate": 2.057816753236874e-06, + "loss": 0.09534336, + "memory(GiB)": 45.75, + "step": 30240, + "train_speed(iter/s)": 0.767979 + }, + { + "epoch": 2.817504655493482, + "grad_norm": 1.359375, + "learning_rate": 2.016594244834391e-06, + "loss": 0.08807804, + "memory(GiB)": 45.75, + "step": 30260, + "train_speed(iter/s)": 0.768021 + }, + { + "epoch": 2.819366852886406, + "grad_norm": 0.9296875, + "learning_rate": 1.9757846069119903e-06, + "loss": 0.09050249, + "memory(GiB)": 45.75, + "step": 30280, + "train_speed(iter/s)": 0.768066 + }, + { + "epoch": 2.82122905027933, + "grad_norm": 1.3203125, + "learning_rate": 1.9353880114283517e-06, + "loss": 0.09604942, + "memory(GiB)": 45.75, + "step": 30300, + "train_speed(iter/s)": 0.768089 + }, + { + "epoch": 2.823091247672253, + "grad_norm": 0.86328125, + "learning_rate": 1.8954046286016936e-06, + "loss": 0.08841636, + "memory(GiB)": 45.75, + "step": 30320, + "train_speed(iter/s)": 0.768128 + }, + { + "epoch": 2.824953445065177, + "grad_norm": 0.96875, + "learning_rate": 1.8558346269090698e-06, + "loss": 0.09765224, + "memory(GiB)": 45.75, + "step": 30340, + "train_speed(iter/s)": 0.768171 + }, + { + "epoch": 2.826815642458101, + "grad_norm": 1.171875, + "learning_rate": 1.816678173085762e-06, + "loss": 0.09548703, + "memory(GiB)": 45.75, + "step": 30360, + "train_speed(iter/s)": 0.768206 + }, + { + "epoch": 2.828677839851024, + "grad_norm": 0.96875, + "learning_rate": 1.7779354321244e-06, + "loss": 0.08739243, + "memory(GiB)": 45.75, + "step": 30380, + "train_speed(iter/s)": 0.768232 + }, + { + "epoch": 2.830540037243948, + "grad_norm": 0.8359375, + "learning_rate": 1.7396065672744211e-06, + "loss": 0.09241841, + "memory(GiB)": 45.75, + "step": 30400, + "train_speed(iter/s)": 0.768247 + }, + { + "epoch": 2.8324022346368714, + "grad_norm": 0.8046875, + "learning_rate": 1.7016917400413001e-06, + "loss": 0.09562107, + "memory(GiB)": 45.75, + "step": 30420, + "train_speed(iter/s)": 0.76828 + }, + { + "epoch": 2.834264432029795, + "grad_norm": 1.3515625, + "learning_rate": 1.6641911101858865e-06, + "loss": 0.0923029, + "memory(GiB)": 45.75, + "step": 30440, + "train_speed(iter/s)": 0.768309 + }, + { + "epoch": 2.8361266294227185, + "grad_norm": 0.890625, + "learning_rate": 1.6271048357237695e-06, + "loss": 0.09228848, + "memory(GiB)": 45.75, + "step": 30460, + "train_speed(iter/s)": 0.768324 + }, + { + "epoch": 2.8379888268156424, + "grad_norm": 0.96875, + "learning_rate": 1.5904330729245243e-06, + "loss": 0.08994877, + "memory(GiB)": 45.75, + "step": 30480, + "train_speed(iter/s)": 0.768352 + }, + { + "epoch": 2.839851024208566, + "grad_norm": 1.140625, + "learning_rate": 1.5541759763111452e-06, + "loss": 0.09694709, + "memory(GiB)": 45.75, + "step": 30500, + "train_speed(iter/s)": 0.768384 + }, + { + "epoch": 2.8417132216014895, + "grad_norm": 1.0546875, + "learning_rate": 1.5183336986593465e-06, + "loss": 0.10389121, + "memory(GiB)": 45.75, + "step": 30520, + "train_speed(iter/s)": 0.768416 + }, + { + "epoch": 2.8435754189944134, + "grad_norm": 1.109375, + "learning_rate": 1.482906390996941e-06, + "loss": 0.08865334, + "memory(GiB)": 45.75, + "step": 30540, + "train_speed(iter/s)": 0.768442 + }, + { + "epoch": 2.845437616387337, + "grad_norm": 0.86328125, + "learning_rate": 1.4478942026031727e-06, + "loss": 0.10575958, + "memory(GiB)": 45.75, + "step": 30560, + "train_speed(iter/s)": 0.768483 + }, + { + "epoch": 2.8472998137802605, + "grad_norm": 0.9453125, + "learning_rate": 1.4132972810080969e-06, + "loss": 0.091862, + "memory(GiB)": 45.75, + "step": 30580, + "train_speed(iter/s)": 0.768508 + }, + { + "epoch": 2.8491620111731844, + "grad_norm": 1.2734375, + "learning_rate": 1.3791157719920122e-06, + "loss": 0.09626628, + "memory(GiB)": 45.75, + "step": 30600, + "train_speed(iter/s)": 0.768524 + }, + { + "epoch": 2.851024208566108, + "grad_norm": 1.640625, + "learning_rate": 1.3453498195847624e-06, + "loss": 0.09501921, + "memory(GiB)": 45.75, + "step": 30620, + "train_speed(iter/s)": 0.768565 + }, + { + "epoch": 2.8528864059590315, + "grad_norm": 1.046875, + "learning_rate": 1.311999566065203e-06, + "loss": 0.10152951, + "memory(GiB)": 45.75, + "step": 30640, + "train_speed(iter/s)": 0.768572 + }, + { + "epoch": 2.8547486033519553, + "grad_norm": 1.5390625, + "learning_rate": 1.2790651519605234e-06, + "loss": 0.09824728, + "memory(GiB)": 45.75, + "step": 30660, + "train_speed(iter/s)": 0.768607 + }, + { + "epoch": 2.856610800744879, + "grad_norm": 1.2421875, + "learning_rate": 1.246546716045749e-06, + "loss": 0.10510842, + "memory(GiB)": 45.75, + "step": 30680, + "train_speed(iter/s)": 0.768631 + }, + { + "epoch": 2.8584729981378025, + "grad_norm": 1.1171875, + "learning_rate": 1.2144443953430617e-06, + "loss": 0.10336709, + "memory(GiB)": 45.75, + "step": 30700, + "train_speed(iter/s)": 0.76866 + }, + { + "epoch": 2.8603351955307263, + "grad_norm": 1.0078125, + "learning_rate": 1.1827583251213026e-06, + "loss": 0.09153603, + "memory(GiB)": 45.75, + "step": 30720, + "train_speed(iter/s)": 0.768687 + }, + { + "epoch": 2.86219739292365, + "grad_norm": 1.0078125, + "learning_rate": 1.1514886388953706e-06, + "loss": 0.08601239, + "memory(GiB)": 45.75, + "step": 30740, + "train_speed(iter/s)": 0.768721 + }, + { + "epoch": 2.8640595903165735, + "grad_norm": 1.0703125, + "learning_rate": 1.1206354684256015e-06, + "loss": 0.09374793, + "memory(GiB)": 45.75, + "step": 30760, + "train_speed(iter/s)": 0.768745 + }, + { + "epoch": 2.8659217877094973, + "grad_norm": 0.7421875, + "learning_rate": 1.0901989437173577e-06, + "loss": 0.09772741, + "memory(GiB)": 45.75, + "step": 30780, + "train_speed(iter/s)": 0.768781 + }, + { + "epoch": 2.8677839851024207, + "grad_norm": 1.0546875, + "learning_rate": 1.0601791930203053e-06, + "loss": 0.10399811, + "memory(GiB)": 45.75, + "step": 30800, + "train_speed(iter/s)": 0.768806 + }, + { + "epoch": 2.8696461824953445, + "grad_norm": 1.1171875, + "learning_rate": 1.0305763428280046e-06, + "loss": 0.10219818, + "memory(GiB)": 45.75, + "step": 30820, + "train_speed(iter/s)": 0.768816 + }, + { + "epoch": 2.871508379888268, + "grad_norm": 0.88671875, + "learning_rate": 1.0013905178773208e-06, + "loss": 0.1023465, + "memory(GiB)": 45.75, + "step": 30840, + "train_speed(iter/s)": 0.768834 + }, + { + "epoch": 2.8733705772811917, + "grad_norm": 0.98046875, + "learning_rate": 9.726218411479137e-07, + "loss": 0.08770326, + "memory(GiB)": 45.75, + "step": 30860, + "train_speed(iter/s)": 0.768871 + }, + { + "epoch": 2.8752327746741155, + "grad_norm": 0.921875, + "learning_rate": 9.442704338617158e-07, + "loss": 0.10444794, + "memory(GiB)": 45.75, + "step": 30880, + "train_speed(iter/s)": 0.768893 + }, + { + "epoch": 2.877094972067039, + "grad_norm": 1.1015625, + "learning_rate": 9.163364154823994e-07, + "loss": 0.0871839, + "memory(GiB)": 45.75, + "step": 30900, + "train_speed(iter/s)": 0.768916 + }, + { + "epoch": 2.8789571694599627, + "grad_norm": 1.1328125, + "learning_rate": 8.888199037149325e-07, + "loss": 0.09597771, + "memory(GiB)": 45.75, + "step": 30920, + "train_speed(iter/s)": 0.768935 + }, + { + "epoch": 2.8808193668528865, + "grad_norm": 0.76953125, + "learning_rate": 8.617210145050014e-07, + "loss": 0.09385652, + "memory(GiB)": 45.75, + "step": 30940, + "train_speed(iter/s)": 0.768951 + }, + { + "epoch": 2.88268156424581, + "grad_norm": 0.6640625, + "learning_rate": 8.350398620386113e-07, + "loss": 0.11076908, + "memory(GiB)": 45.75, + "step": 30960, + "train_speed(iter/s)": 0.768962 + }, + { + "epoch": 2.8845437616387337, + "grad_norm": 0.98046875, + "learning_rate": 8.087765587415197e-07, + "loss": 0.08995301, + "memory(GiB)": 45.75, + "step": 30980, + "train_speed(iter/s)": 0.768978 + }, + { + "epoch": 2.8864059590316575, + "grad_norm": 0.81640625, + "learning_rate": 7.82931215278826e-07, + "loss": 0.10020612, + "memory(GiB)": 45.75, + "step": 31000, + "train_speed(iter/s)": 0.768993 + }, + { + "epoch": 2.888268156424581, + "grad_norm": 1.0, + "learning_rate": 7.575039405544715e-07, + "loss": 0.08659879, + "memory(GiB)": 45.75, + "step": 31020, + "train_speed(iter/s)": 0.76903 + }, + { + "epoch": 2.8901303538175047, + "grad_norm": 0.96875, + "learning_rate": 7.324948417107957e-07, + "loss": 0.08481673, + "memory(GiB)": 45.75, + "step": 31040, + "train_speed(iter/s)": 0.769054 + }, + { + "epoch": 2.8919925512104285, + "grad_norm": 1.0390625, + "learning_rate": 7.07904024128081e-07, + "loss": 0.09589967, + "memory(GiB)": 45.75, + "step": 31060, + "train_speed(iter/s)": 0.769068 + }, + { + "epoch": 2.893854748603352, + "grad_norm": 0.984375, + "learning_rate": 6.837315914240972e-07, + "loss": 0.08965684, + "memory(GiB)": 45.75, + "step": 31080, + "train_speed(iter/s)": 0.769106 + }, + { + "epoch": 2.8957169459962757, + "grad_norm": 0.89453125, + "learning_rate": 6.59977645453691e-07, + "loss": 0.09921396, + "memory(GiB)": 45.75, + "step": 31100, + "train_speed(iter/s)": 0.769127 + }, + { + "epoch": 2.8975791433891995, + "grad_norm": 0.70703125, + "learning_rate": 6.366422863083199e-07, + "loss": 0.09246094, + "memory(GiB)": 45.75, + "step": 31120, + "train_speed(iter/s)": 0.769148 + }, + { + "epoch": 2.899441340782123, + "grad_norm": 1.0078125, + "learning_rate": 6.137256123156631e-07, + "loss": 0.09109651, + "memory(GiB)": 45.75, + "step": 31140, + "train_speed(iter/s)": 0.769182 + }, + { + "epoch": 2.9013035381750467, + "grad_norm": 0.703125, + "learning_rate": 5.912277200392113e-07, + "loss": 0.08994749, + "memory(GiB)": 45.75, + "step": 31160, + "train_speed(iter/s)": 0.769206 + }, + { + "epoch": 2.90316573556797, + "grad_norm": 0.88671875, + "learning_rate": 5.69148704277811e-07, + "loss": 0.08762764, + "memory(GiB)": 45.75, + "step": 31180, + "train_speed(iter/s)": 0.769234 + }, + { + "epoch": 2.905027932960894, + "grad_norm": 1.234375, + "learning_rate": 5.474886580653315e-07, + "loss": 0.0921519, + "memory(GiB)": 45.75, + "step": 31200, + "train_speed(iter/s)": 0.769261 + }, + { + "epoch": 2.9068901303538173, + "grad_norm": 0.9296875, + "learning_rate": 5.262476726702215e-07, + "loss": 0.09754463, + "memory(GiB)": 45.75, + "step": 31220, + "train_speed(iter/s)": 0.769276 + }, + { + "epoch": 2.908752327746741, + "grad_norm": 0.9921875, + "learning_rate": 5.054258375951415e-07, + "loss": 0.08906224, + "memory(GiB)": 45.75, + "step": 31240, + "train_speed(iter/s)": 0.769297 + }, + { + "epoch": 2.910614525139665, + "grad_norm": 1.078125, + "learning_rate": 4.850232405765875e-07, + "loss": 0.09115668, + "memory(GiB)": 45.75, + "step": 31260, + "train_speed(iter/s)": 0.769324 + }, + { + "epoch": 2.9124767225325883, + "grad_norm": 1.03125, + "learning_rate": 4.6503996758452407e-07, + "loss": 0.08493429, + "memory(GiB)": 45.75, + "step": 31280, + "train_speed(iter/s)": 0.76933 + }, + { + "epoch": 2.914338919925512, + "grad_norm": 0.9765625, + "learning_rate": 4.454761028220289e-07, + "loss": 0.09017068, + "memory(GiB)": 45.75, + "step": 31300, + "train_speed(iter/s)": 0.769354 + }, + { + "epoch": 2.916201117318436, + "grad_norm": 1.0390625, + "learning_rate": 4.263317287249158e-07, + "loss": 0.08491018, + "memory(GiB)": 45.75, + "step": 31320, + "train_speed(iter/s)": 0.769377 + }, + { + "epoch": 2.9180633147113593, + "grad_norm": 1.0390625, + "learning_rate": 4.0760692596140124e-07, + "loss": 0.08563709, + "memory(GiB)": 45.75, + "step": 31340, + "train_speed(iter/s)": 0.769395 + }, + { + "epoch": 2.919925512104283, + "grad_norm": 1.0078125, + "learning_rate": 3.8930177343177165e-07, + "loss": 0.08611808, + "memory(GiB)": 45.75, + "step": 31360, + "train_speed(iter/s)": 0.769403 + }, + { + "epoch": 2.921787709497207, + "grad_norm": 1.15625, + "learning_rate": 3.714163482680388e-07, + "loss": 0.09184546, + "memory(GiB)": 45.75, + "step": 31380, + "train_speed(iter/s)": 0.76942 + }, + { + "epoch": 2.9236499068901303, + "grad_norm": 1.03125, + "learning_rate": 3.5395072583361833e-07, + "loss": 0.09664055, + "memory(GiB)": 45.75, + "step": 31400, + "train_speed(iter/s)": 0.769447 + }, + { + "epoch": 2.925512104283054, + "grad_norm": 1.171875, + "learning_rate": 3.369049797230073e-07, + "loss": 0.10272588, + "memory(GiB)": 45.75, + "step": 31420, + "train_speed(iter/s)": 0.769475 + }, + { + "epoch": 2.927374301675978, + "grad_norm": 0.8125, + "learning_rate": 3.2027918176148474e-07, + "loss": 0.08928733, + "memory(GiB)": 45.75, + "step": 31440, + "train_speed(iter/s)": 0.769516 + }, + { + "epoch": 2.9292364990689013, + "grad_norm": 1.015625, + "learning_rate": 3.040734020048008e-07, + "loss": 0.10981673, + "memory(GiB)": 45.75, + "step": 31460, + "train_speed(iter/s)": 0.769512 + }, + { + "epoch": 2.931098696461825, + "grad_norm": 1.078125, + "learning_rate": 2.88287708738888e-07, + "loss": 0.0986732, + "memory(GiB)": 45.75, + "step": 31480, + "train_speed(iter/s)": 0.769534 + }, + { + "epoch": 2.9329608938547485, + "grad_norm": 1.03125, + "learning_rate": 2.7292216847957244e-07, + "loss": 0.0985683, + "memory(GiB)": 45.75, + "step": 31500, + "train_speed(iter/s)": 0.769578 + }, + { + "epoch": 2.9348230912476723, + "grad_norm": 1.171875, + "learning_rate": 2.579768459722853e-07, + "loss": 0.09273306, + "memory(GiB)": 45.75, + "step": 31520, + "train_speed(iter/s)": 0.769616 + }, + { + "epoch": 2.9366852886405956, + "grad_norm": 1.203125, + "learning_rate": 2.434518041917966e-07, + "loss": 0.10807085, + "memory(GiB)": 45.75, + "step": 31540, + "train_speed(iter/s)": 0.769631 + }, + { + "epoch": 2.9385474860335195, + "grad_norm": 0.8984375, + "learning_rate": 2.2934710434194818e-07, + "loss": 0.10285647, + "memory(GiB)": 45.75, + "step": 31560, + "train_speed(iter/s)": 0.769659 + }, + { + "epoch": 2.9404096834264433, + "grad_norm": 0.9296875, + "learning_rate": 2.1566280585539888e-07, + "loss": 0.09199415, + "memory(GiB)": 45.75, + "step": 31580, + "train_speed(iter/s)": 0.76968 + }, + { + "epoch": 2.9422718808193666, + "grad_norm": 1.265625, + "learning_rate": 2.0239896639336898e-07, + "loss": 0.09799184, + "memory(GiB)": 45.75, + "step": 31600, + "train_speed(iter/s)": 0.769719 + }, + { + "epoch": 2.9441340782122905, + "grad_norm": 0.98828125, + "learning_rate": 1.8955564184539587e-07, + "loss": 0.0983022, + "memory(GiB)": 45.75, + "step": 31620, + "train_speed(iter/s)": 0.769745 + }, + { + "epoch": 2.9459962756052143, + "grad_norm": 0.8671875, + "learning_rate": 1.771328863291233e-07, + "loss": 0.09433325, + "memory(GiB)": 45.75, + "step": 31640, + "train_speed(iter/s)": 0.769762 + }, + { + "epoch": 2.9478584729981376, + "grad_norm": 0.875, + "learning_rate": 1.6513075219001252e-07, + "loss": 0.09435837, + "memory(GiB)": 45.75, + "step": 31660, + "train_speed(iter/s)": 0.769781 + }, + { + "epoch": 2.9497206703910615, + "grad_norm": 0.96875, + "learning_rate": 1.5354929000122032e-07, + "loss": 0.09303812, + "memory(GiB)": 45.75, + "step": 31680, + "train_speed(iter/s)": 0.769783 + }, + { + "epoch": 2.9515828677839853, + "grad_norm": 1.125, + "learning_rate": 1.423885485632659e-07, + "loss": 0.09572449, + "memory(GiB)": 45.75, + "step": 31700, + "train_speed(iter/s)": 0.769797 + }, + { + "epoch": 2.9534450651769086, + "grad_norm": 1.53125, + "learning_rate": 1.316485749039309e-07, + "loss": 0.10269543, + "memory(GiB)": 45.75, + "step": 31720, + "train_speed(iter/s)": 0.769825 + }, + { + "epoch": 2.9553072625698324, + "grad_norm": 1.203125, + "learning_rate": 1.2132941427799304e-07, + "loss": 0.0916387, + "memory(GiB)": 45.75, + "step": 31740, + "train_speed(iter/s)": 0.769837 + }, + { + "epoch": 2.9571694599627563, + "grad_norm": 1.0078125, + "learning_rate": 1.1143111016707063e-07, + "loss": 0.09537607, + "memory(GiB)": 45.75, + "step": 31760, + "train_speed(iter/s)": 0.769875 + }, + { + "epoch": 2.9590316573556796, + "grad_norm": 1.03125, + "learning_rate": 1.0195370427943385e-07, + "loss": 0.10331657, + "memory(GiB)": 45.75, + "step": 31780, + "train_speed(iter/s)": 0.76991 + }, + { + "epoch": 2.9608938547486034, + "grad_norm": 1.2421875, + "learning_rate": 9.289723654980486e-08, + "loss": 0.0982655, + "memory(GiB)": 45.75, + "step": 31800, + "train_speed(iter/s)": 0.769926 + }, + { + "epoch": 2.9627560521415273, + "grad_norm": 0.984375, + "learning_rate": 8.426174513922469e-08, + "loss": 0.09388782, + "memory(GiB)": 45.75, + "step": 31820, + "train_speed(iter/s)": 0.769945 + }, + { + "epoch": 2.9646182495344506, + "grad_norm": 1.0, + "learning_rate": 7.604726643486438e-08, + "loss": 0.09825534, + "memory(GiB)": 45.75, + "step": 31840, + "train_speed(iter/s)": 0.769959 + }, + { + "epoch": 2.9664804469273744, + "grad_norm": 1.0546875, + "learning_rate": 6.82538350498918e-08, + "loss": 0.10639879, + "memory(GiB)": 45.75, + "step": 31860, + "train_speed(iter/s)": 0.76997 + }, + { + "epoch": 2.968342644320298, + "grad_norm": 1.1171875, + "learning_rate": 6.088148382331627e-08, + "loss": 0.09718776, + "memory(GiB)": 45.75, + "step": 31880, + "train_speed(iter/s)": 0.769995 + }, + { + "epoch": 2.9702048417132216, + "grad_norm": 1.140625, + "learning_rate": 5.393024381984413e-08, + "loss": 0.08618087, + "memory(GiB)": 45.75, + "step": 31900, + "train_speed(iter/s)": 0.770024 + }, + { + "epoch": 2.972067039106145, + "grad_norm": 0.7890625, + "learning_rate": 4.74001443297678e-08, + "loss": 0.09072049, + "memory(GiB)": 45.75, + "step": 31920, + "train_speed(iter/s)": 0.770048 + }, + { + "epoch": 2.973929236499069, + "grad_norm": 1.3046875, + "learning_rate": 4.129121286879922e-08, + "loss": 0.09811844, + "memory(GiB)": 45.75, + "step": 31940, + "train_speed(iter/s)": 0.770087 + }, + { + "epoch": 2.9757914338919926, + "grad_norm": 0.8828125, + "learning_rate": 3.5603475178036526e-08, + "loss": 0.08870932, + "memory(GiB)": 45.75, + "step": 31960, + "train_speed(iter/s)": 0.770105 + }, + { + "epoch": 2.977653631284916, + "grad_norm": 1.21875, + "learning_rate": 3.033695522375313e-08, + "loss": 0.09320135, + "memory(GiB)": 45.75, + "step": 31980, + "train_speed(iter/s)": 0.770131 + }, + { + "epoch": 2.97951582867784, + "grad_norm": 0.94921875, + "learning_rate": 2.5491675197375498e-08, + "loss": 0.10496864, + "memory(GiB)": 45.75, + "step": 32000, + "train_speed(iter/s)": 0.770166 + }, + { + "epoch": 2.97951582867784, + "eval_loss": 0.3874885141849518, + "eval_runtime": 76.9471, + "eval_samples_per_second": 180.449, + "eval_steps_per_second": 1.417, + "step": 32000 + }, + { + "epoch": 2.9813780260707636, + "grad_norm": 1.1328125, + "learning_rate": 2.106765551534995e-08, + "loss": 0.09619021, + "memory(GiB)": 45.75, + "step": 32020, + "train_speed(iter/s)": 0.768206 + }, + { + "epoch": 2.983240223463687, + "grad_norm": 0.75, + "learning_rate": 1.7064914819064915e-08, + "loss": 0.08765938, + "memory(GiB)": 45.75, + "step": 32040, + "train_speed(iter/s)": 0.768239 + }, + { + "epoch": 2.985102420856611, + "grad_norm": 1.171875, + "learning_rate": 1.3483469974773233e-08, + "loss": 0.1008575, + "memory(GiB)": 45.75, + "step": 32060, + "train_speed(iter/s)": 0.768255 + }, + { + "epoch": 2.9869646182495346, + "grad_norm": 1.140625, + "learning_rate": 1.0323336073536638e-08, + "loss": 0.10057193, + "memory(GiB)": 45.75, + "step": 32080, + "train_speed(iter/s)": 0.768274 + }, + { + "epoch": 2.988826815642458, + "grad_norm": 1.1484375, + "learning_rate": 7.584526431148043e-09, + "loss": 0.08665556, + "memory(GiB)": 45.75, + "step": 32100, + "train_speed(iter/s)": 0.768288 + }, + { + "epoch": 2.990689013035382, + "grad_norm": 0.91796875, + "learning_rate": 5.267052588053822e-09, + "loss": 0.09432332, + "memory(GiB)": 45.75, + "step": 32120, + "train_speed(iter/s)": 0.768312 + }, + { + "epoch": 2.9925512104283056, + "grad_norm": 0.90625, + "learning_rate": 3.370924309342716e-09, + "loss": 0.10474067, + "memory(GiB)": 45.75, + "step": 32140, + "train_speed(iter/s)": 0.768318 + }, + { + "epoch": 2.994413407821229, + "grad_norm": 1.203125, + "learning_rate": 1.8961495847014122e-09, + "loss": 0.08654416, + "memory(GiB)": 45.75, + "step": 32160, + "train_speed(iter/s)": 0.768345 + }, + { + "epoch": 2.996275605214153, + "grad_norm": 1.15625, + "learning_rate": 8.427346283479409e-10, + "loss": 0.1108827, + "memory(GiB)": 45.75, + "step": 32180, + "train_speed(iter/s)": 0.768396 + }, + { + "epoch": 2.998137802607076, + "grad_norm": 1.1015625, + "learning_rate": 2.1068387902056786e-10, + "loss": 0.09056249, + "memory(GiB)": 45.75, + "step": 32200, + "train_speed(iter/s)": 0.768395 + }, + { + "epoch": 3.0, + "grad_norm": 1.0625, + "learning_rate": 0.0, + "loss": 0.08908014, + "memory(GiB)": 45.75, + "step": 32220, + "train_speed(iter/s)": 0.76839 + }, + { + "epoch": 3.0, + "eval_loss": 0.3871558606624603, + "eval_runtime": 76.4163, + "eval_samples_per_second": 181.702, + "eval_steps_per_second": 1.426, + "step": 32220 + } + ], + "logging_steps": 20, + "max_steps": 32220, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 2000, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 4.14434535928968e+18, + "train_batch_size": 16, + "trial_name": null, + "trial_params": null +}