| { | |
| "best_metric": 12.417525291442871, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-200", | |
| "epoch": 0.12961762799740764, | |
| "eval_steps": 50, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0006480881399870382, | |
| "grad_norm": 0.006569863297045231, | |
| "learning_rate": 1e-05, | |
| "loss": 12.457, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0006480881399870382, | |
| "eval_loss": 12.451666831970215, | |
| "eval_runtime": 25.7675, | |
| "eval_samples_per_second": 100.863, | |
| "eval_steps_per_second": 25.226, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0012961762799740765, | |
| "grad_norm": 0.00675651291385293, | |
| "learning_rate": 2e-05, | |
| "loss": 12.4568, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0019442644199611147, | |
| "grad_norm": 0.0072723510675132275, | |
| "learning_rate": 3e-05, | |
| "loss": 12.4595, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.002592352559948153, | |
| "grad_norm": 0.008350463584065437, | |
| "learning_rate": 4e-05, | |
| "loss": 12.442, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0032404406999351912, | |
| "grad_norm": 0.0108029218390584, | |
| "learning_rate": 5e-05, | |
| "loss": 12.4407, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0038885288399222295, | |
| "grad_norm": 0.007661391980946064, | |
| "learning_rate": 6e-05, | |
| "loss": 12.4444, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.004536616979909268, | |
| "grad_norm": 0.007625883910804987, | |
| "learning_rate": 7e-05, | |
| "loss": 12.4525, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.005184705119896306, | |
| "grad_norm": 0.008109038695693016, | |
| "learning_rate": 8e-05, | |
| "loss": 12.4566, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.005832793259883344, | |
| "grad_norm": 0.008637198247015476, | |
| "learning_rate": 9e-05, | |
| "loss": 12.454, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0064808813998703824, | |
| "grad_norm": 0.008731589652597904, | |
| "learning_rate": 0.0001, | |
| "loss": 12.4523, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.007128969539857421, | |
| "grad_norm": 0.008254251442849636, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 12.4491, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.007777057679844459, | |
| "grad_norm": 0.011887218803167343, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 12.4472, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.008425145819831496, | |
| "grad_norm": 0.00914101954549551, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 12.4525, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.009073233959818535, | |
| "grad_norm": 0.010923495516180992, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 12.4539, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.009721322099805573, | |
| "grad_norm": 0.008453849703073502, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 12.454, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.010369410239792612, | |
| "grad_norm": 0.011008517816662788, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 12.4486, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.01101749837977965, | |
| "grad_norm": 0.011536368168890476, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 12.4604, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.011665586519766688, | |
| "grad_norm": 0.013136766850948334, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 12.4555, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.012313674659753726, | |
| "grad_norm": 0.011392605490982533, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 12.4526, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.012961762799740765, | |
| "grad_norm": 0.013982742093503475, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 12.4603, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.013609850939727802, | |
| "grad_norm": 0.011519525200128555, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 12.4564, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.014257939079714841, | |
| "grad_norm": 0.012463117018342018, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 12.4544, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.014906027219701879, | |
| "grad_norm": 0.014147116802632809, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 12.4576, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.015554115359688918, | |
| "grad_norm": 0.014245379716157913, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 12.458, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.016202203499675955, | |
| "grad_norm": 0.015419174917042255, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 12.4524, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.016850291639662993, | |
| "grad_norm": 0.015862194821238518, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 12.4394, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.017498379779650033, | |
| "grad_norm": 0.01722586899995804, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 12.4497, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.01814646791963707, | |
| "grad_norm": 0.021126728504896164, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 12.4515, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.018794556059624108, | |
| "grad_norm": 0.017525460571050644, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 12.456, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.019442644199611146, | |
| "grad_norm": 0.02138211391866207, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 12.4504, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.020090732339598186, | |
| "grad_norm": 0.021518006920814514, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 12.4428, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.020738820479585224, | |
| "grad_norm": 0.02515367604792118, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 12.4526, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.02138690861957226, | |
| "grad_norm": 0.02069803886115551, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 12.4571, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.0220349967595593, | |
| "grad_norm": 0.023277193307876587, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 12.4441, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.02268308489954634, | |
| "grad_norm": 0.019151320680975914, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 12.4503, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.023331173039533377, | |
| "grad_norm": 0.02539009414613247, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 12.4532, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.023979261179520414, | |
| "grad_norm": 0.028782585635781288, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 12.4455, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.02462734931950745, | |
| "grad_norm": 0.027810685336589813, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 12.4534, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.025275437459494492, | |
| "grad_norm": 0.0357164666056633, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 12.4513, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.02592352559948153, | |
| "grad_norm": 0.03279275819659233, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 12.4486, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.026571613739468567, | |
| "grad_norm": 0.03006412461400032, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 12.4498, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.027219701879455604, | |
| "grad_norm": 0.032223720103502274, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 12.4521, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.027867790019442645, | |
| "grad_norm": 0.032283008098602295, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 12.4564, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.028515878159429683, | |
| "grad_norm": 0.03774842619895935, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 12.452, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.02916396629941672, | |
| "grad_norm": 0.041357334703207016, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 12.4484, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.029812054439403757, | |
| "grad_norm": 0.035570479929447174, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 12.4531, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.0304601425793908, | |
| "grad_norm": 0.037538446485996246, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 12.4423, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.031108230719377836, | |
| "grad_norm": 0.037077080458402634, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 12.4469, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.031756318859364877, | |
| "grad_norm": 0.04468686506152153, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 12.4471, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.03240440699935191, | |
| "grad_norm": 0.05200482904911041, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 12.4515, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03240440699935191, | |
| "eval_loss": 12.447880744934082, | |
| "eval_runtime": 25.6247, | |
| "eval_samples_per_second": 101.425, | |
| "eval_steps_per_second": 25.366, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03305249513933895, | |
| "grad_norm": 0.031452253460884094, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 12.4558, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.033700583279325985, | |
| "grad_norm": 0.034291695803403854, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 12.4501, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.034348671419313026, | |
| "grad_norm": 0.041694141924381256, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 12.4471, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.03499675955930007, | |
| "grad_norm": 0.03913399949669838, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 12.4289, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.0356448476992871, | |
| "grad_norm": 0.04792981967329979, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 12.4314, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.03629293583927414, | |
| "grad_norm": 0.04870151728391647, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 12.4378, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.03694102397926118, | |
| "grad_norm": 0.04969751834869385, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 12.4418, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.037589112119248216, | |
| "grad_norm": 0.05134296417236328, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 12.4507, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.03823720025923526, | |
| "grad_norm": 0.05872856453061104, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 12.4442, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.03888528839922229, | |
| "grad_norm": 0.05536626651883125, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 12.4463, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.03953337653920933, | |
| "grad_norm": 0.07720275968313217, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 12.4393, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.04018146467919637, | |
| "grad_norm": 0.0935516208410263, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 12.437, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.04082955281918341, | |
| "grad_norm": 0.06796793639659882, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 12.4422, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.04147764095917045, | |
| "grad_norm": 0.050208643078804016, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 12.4564, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.04212572909915749, | |
| "grad_norm": 0.05703168734908104, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 12.4502, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.04277381723914452, | |
| "grad_norm": 0.06285442411899567, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 12.4449, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.04342190537913156, | |
| "grad_norm": 0.06523274630308151, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 12.4389, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.0440699935191186, | |
| "grad_norm": 0.06631460040807724, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 12.4438, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.04471808165910564, | |
| "grad_norm": 0.0676867812871933, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 12.4423, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.04536616979909268, | |
| "grad_norm": 0.0614280179142952, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 12.4473, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.04601425793907971, | |
| "grad_norm": 0.07548827677965164, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 12.4455, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.046662346079066754, | |
| "grad_norm": 0.0685398280620575, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 12.4387, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.047310434219053794, | |
| "grad_norm": 0.07349091023206711, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 12.4393, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.04795852235904083, | |
| "grad_norm": 0.0725373849272728, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 12.4429, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.04860661049902787, | |
| "grad_norm": 0.08269369602203369, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 12.4424, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.0492546986390149, | |
| "grad_norm": 0.09406210482120514, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 12.445, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.049902786779001944, | |
| "grad_norm": 0.08032267540693283, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 12.4465, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.050550874918988985, | |
| "grad_norm": 0.06973438709974289, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 12.4379, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.05119896305897602, | |
| "grad_norm": 0.07986024767160416, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 12.4467, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.05184705119896306, | |
| "grad_norm": 0.09152371436357498, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 12.4432, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0524951393389501, | |
| "grad_norm": 0.08582999557256699, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 12.4425, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.053143227478937134, | |
| "grad_norm": 0.08873876184225082, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 12.4345, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.053791315618924175, | |
| "grad_norm": 0.08987562358379364, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 12.4335, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.05443940375891121, | |
| "grad_norm": 0.09282175451517105, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 12.444, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.05508749189889825, | |
| "grad_norm": 0.09211423993110657, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 12.4386, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.05573558003888529, | |
| "grad_norm": 0.09282619506120682, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 12.4366, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.056383668178872325, | |
| "grad_norm": 0.09386464208364487, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 12.4353, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.057031756318859365, | |
| "grad_norm": 0.11244712024927139, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 12.4347, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.057679844458846406, | |
| "grad_norm": 0.11355287581682205, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 12.4393, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.05832793259883344, | |
| "grad_norm": 0.09205302596092224, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 12.4333, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.05897602073882048, | |
| "grad_norm": 0.10400447994470596, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 12.431, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.059624108878807515, | |
| "grad_norm": 0.10883035510778427, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 12.4312, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.060272197018794556, | |
| "grad_norm": 0.0981161817908287, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 12.4276, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.0609202851587816, | |
| "grad_norm": 0.10417820513248444, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 12.431, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.06156837329876863, | |
| "grad_norm": 0.08984676003456116, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 12.4349, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.06221646143875567, | |
| "grad_norm": 0.10311862826347351, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 12.4285, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.06286454957874271, | |
| "grad_norm": 0.1088971346616745, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 12.427, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.06351263771872975, | |
| "grad_norm": 0.10161225497722626, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 12.4385, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.06416072585871678, | |
| "grad_norm": 0.11141715198755264, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 12.4309, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.06480881399870382, | |
| "grad_norm": 0.09958706051111221, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 12.4312, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06480881399870382, | |
| "eval_loss": 12.429600715637207, | |
| "eval_runtime": 25.6262, | |
| "eval_samples_per_second": 101.42, | |
| "eval_steps_per_second": 25.365, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06545690213869086, | |
| "grad_norm": 0.07751390337944031, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 12.4466, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.0661049902786779, | |
| "grad_norm": 0.08995895832777023, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 12.4475, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.06675307841866494, | |
| "grad_norm": 0.07873629033565521, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 12.4403, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.06740116655865197, | |
| "grad_norm": 0.0779220312833786, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 12.4203, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.06804925469863901, | |
| "grad_norm": 0.08301982283592224, | |
| "learning_rate": 5e-05, | |
| "loss": 12.4321, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.06869734283862605, | |
| "grad_norm": 0.1038648933172226, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 12.4148, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.06934543097861309, | |
| "grad_norm": 0.09731864929199219, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 12.4237, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.06999351911860013, | |
| "grad_norm": 0.09959754347801208, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 12.436, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.07064160725858717, | |
| "grad_norm": 0.10399919748306274, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 12.4304, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.0712896953985742, | |
| "grad_norm": 0.11904186755418777, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 12.4322, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.07193778353856124, | |
| "grad_norm": 0.11899296939373016, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 12.417, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.07258587167854828, | |
| "grad_norm": 0.11149235814809799, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 12.4146, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.07323395981853532, | |
| "grad_norm": 0.11097011715173721, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 12.4287, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.07388204795852236, | |
| "grad_norm": 0.10374613106250763, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 12.4261, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.07453013609850939, | |
| "grad_norm": 0.09058938920497894, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 12.4274, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.07517822423849643, | |
| "grad_norm": 0.09312349557876587, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 12.4105, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.07582631237848347, | |
| "grad_norm": 0.09642358124256134, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 12.4338, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.07647440051847051, | |
| "grad_norm": 0.08863256871700287, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 12.4312, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.07712248865845756, | |
| "grad_norm": 0.11147978156805038, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 12.4325, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.07777057679844458, | |
| "grad_norm": 0.09696000069379807, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 12.4219, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.07841866493843162, | |
| "grad_norm": 0.09084559977054596, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 12.427, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.07906675307841866, | |
| "grad_norm": 0.09775815904140472, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 12.4302, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.0797148412184057, | |
| "grad_norm": 0.0837833359837532, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 12.4261, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.08036292935839275, | |
| "grad_norm": 0.09297660738229752, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 12.4277, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.08101101749837979, | |
| "grad_norm": 0.1246127262711525, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 12.4315, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.08165910563836681, | |
| "grad_norm": 0.08419708907604218, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 12.4205, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.08230719377835385, | |
| "grad_norm": 0.10026169568300247, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 12.4226, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.0829552819183409, | |
| "grad_norm": 0.09828945994377136, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 12.4341, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.08360337005832794, | |
| "grad_norm": 0.11857884377241135, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 12.4262, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.08425145819831498, | |
| "grad_norm": 0.12032639235258102, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 12.416, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.084899546338302, | |
| "grad_norm": 0.11306501179933548, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 12.4165, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.08554763447828904, | |
| "grad_norm": 0.1219179555773735, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 12.4186, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.08619572261827609, | |
| "grad_norm": 0.09922118484973907, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 12.4173, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.08684381075826313, | |
| "grad_norm": 0.11537378281354904, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 12.4318, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.08749189889825017, | |
| "grad_norm": 0.08591119945049286, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 12.4141, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.0881399870382372, | |
| "grad_norm": 0.10934874415397644, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 12.4227, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.08878807517822424, | |
| "grad_norm": 0.12119291722774506, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 12.4217, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.08943616331821128, | |
| "grad_norm": 0.07892826944589615, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 12.4167, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.09008425145819832, | |
| "grad_norm": 0.1038740873336792, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 12.4115, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.09073233959818536, | |
| "grad_norm": 0.08994939923286438, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 12.4271, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.09138042773817238, | |
| "grad_norm": 0.09321318566799164, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 12.4191, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.09202851587815943, | |
| "grad_norm": 0.0827326849102974, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 12.4233, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.09267660401814647, | |
| "grad_norm": 0.10651781409978867, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 12.424, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.09332469215813351, | |
| "grad_norm": 0.09245728701353073, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 12.426, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.09397278029812055, | |
| "grad_norm": 0.10340756922960281, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 12.4263, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.09462086843810759, | |
| "grad_norm": 0.08328746259212494, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 12.4199, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.09526895657809462, | |
| "grad_norm": 0.08074817061424255, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 12.422, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.09591704471808166, | |
| "grad_norm": 0.09953344613313675, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 12.4178, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.0965651328580687, | |
| "grad_norm": 0.09803630411624908, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 12.4225, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.09721322099805574, | |
| "grad_norm": 0.09342999756336212, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 12.4217, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09721322099805574, | |
| "eval_loss": 12.418988227844238, | |
| "eval_runtime": 25.5459, | |
| "eval_samples_per_second": 101.738, | |
| "eval_steps_per_second": 25.444, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09786130913804278, | |
| "grad_norm": 0.07489677518606186, | |
| "learning_rate": 1.553232954407171e-05, | |
| "loss": 12.4404, | |
| "step": 151 | |
| }, | |
| { | |
| "epoch": 0.0985093972780298, | |
| "grad_norm": 0.07785847038030624, | |
| "learning_rate": 1.4938160786375572e-05, | |
| "loss": 12.4306, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 0.09915748541801685, | |
| "grad_norm": 0.07153536379337311, | |
| "learning_rate": 1.435357758543015e-05, | |
| "loss": 12.4072, | |
| "step": 153 | |
| }, | |
| { | |
| "epoch": 0.09980557355800389, | |
| "grad_norm": 0.08717621862888336, | |
| "learning_rate": 1.3778739760445552e-05, | |
| "loss": 12.4147, | |
| "step": 154 | |
| }, | |
| { | |
| "epoch": 0.10045366169799093, | |
| "grad_norm": 0.07147811353206635, | |
| "learning_rate": 1.3213804466343421e-05, | |
| "loss": 12.4082, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.10110174983797797, | |
| "grad_norm": 0.1026967242360115, | |
| "learning_rate": 1.2658926150792322e-05, | |
| "loss": 12.4237, | |
| "step": 156 | |
| }, | |
| { | |
| "epoch": 0.101749837977965, | |
| "grad_norm": 0.09502165019512177, | |
| "learning_rate": 1.2114256511983274e-05, | |
| "loss": 12.4178, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 0.10239792611795204, | |
| "grad_norm": 0.0957978218793869, | |
| "learning_rate": 1.157994445715706e-05, | |
| "loss": 12.4239, | |
| "step": 158 | |
| }, | |
| { | |
| "epoch": 0.10304601425793908, | |
| "grad_norm": 0.08354347944259644, | |
| "learning_rate": 1.1056136061894384e-05, | |
| "loss": 12.4217, | |
| "step": 159 | |
| }, | |
| { | |
| "epoch": 0.10369410239792612, | |
| "grad_norm": 0.09865592420101166, | |
| "learning_rate": 1.0542974530180327e-05, | |
| "loss": 12.4231, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.10434219053791316, | |
| "grad_norm": 0.13387469947338104, | |
| "learning_rate": 1.0040600155253765e-05, | |
| "loss": 12.4159, | |
| "step": 161 | |
| }, | |
| { | |
| "epoch": 0.1049902786779002, | |
| "grad_norm": 0.10183194279670715, | |
| "learning_rate": 9.549150281252633e-06, | |
| "loss": 12.4149, | |
| "step": 162 | |
| }, | |
| { | |
| "epoch": 0.10563836681788723, | |
| "grad_norm": 0.11247194558382034, | |
| "learning_rate": 9.068759265665384e-06, | |
| "loss": 12.4209, | |
| "step": 163 | |
| }, | |
| { | |
| "epoch": 0.10628645495787427, | |
| "grad_norm": 0.11142536997795105, | |
| "learning_rate": 8.599558442598998e-06, | |
| "loss": 12.4204, | |
| "step": 164 | |
| }, | |
| { | |
| "epoch": 0.10693454309786131, | |
| "grad_norm": 0.09879989176988602, | |
| "learning_rate": 8.141676086873572e-06, | |
| "loss": 12.4134, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.10758263123784835, | |
| "grad_norm": 0.07827907055616379, | |
| "learning_rate": 7.695237378953223e-06, | |
| "loss": 12.419, | |
| "step": 166 | |
| }, | |
| { | |
| "epoch": 0.10823071937783539, | |
| "grad_norm": 0.09458613395690918, | |
| "learning_rate": 7.260364370723044e-06, | |
| "loss": 12.4172, | |
| "step": 167 | |
| }, | |
| { | |
| "epoch": 0.10887880751782242, | |
| "grad_norm": 0.10593276470899582, | |
| "learning_rate": 6.837175952121306e-06, | |
| "loss": 12.4185, | |
| "step": 168 | |
| }, | |
| { | |
| "epoch": 0.10952689565780946, | |
| "grad_norm": 0.09929781407117844, | |
| "learning_rate": 6.425787818636131e-06, | |
| "loss": 12.4202, | |
| "step": 169 | |
| }, | |
| { | |
| "epoch": 0.1101749837977965, | |
| "grad_norm": 0.10400168597698212, | |
| "learning_rate": 6.026312439675552e-06, | |
| "loss": 12.4083, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.11082307193778354, | |
| "grad_norm": 0.07882007956504822, | |
| "learning_rate": 5.6388590278194096e-06, | |
| "loss": 12.4225, | |
| "step": 171 | |
| }, | |
| { | |
| "epoch": 0.11147116007777058, | |
| "grad_norm": 0.07449810206890106, | |
| "learning_rate": 5.263533508961827e-06, | |
| "loss": 12.4193, | |
| "step": 172 | |
| }, | |
| { | |
| "epoch": 0.11211924821775761, | |
| "grad_norm": 0.10355754941701889, | |
| "learning_rate": 4.900438493352055e-06, | |
| "loss": 12.4231, | |
| "step": 173 | |
| }, | |
| { | |
| "epoch": 0.11276733635774465, | |
| "grad_norm": 0.09946378320455551, | |
| "learning_rate": 4.549673247541875e-06, | |
| "loss": 12.4171, | |
| "step": 174 | |
| }, | |
| { | |
| "epoch": 0.11341542449773169, | |
| "grad_norm": 0.07985047250986099, | |
| "learning_rate": 4.2113336672471245e-06, | |
| "loss": 12.4218, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.11406351263771873, | |
| "grad_norm": 0.08866812288761139, | |
| "learning_rate": 3.885512251130763e-06, | |
| "loss": 12.4223, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 0.11471160077770577, | |
| "grad_norm": 0.09334906935691833, | |
| "learning_rate": 3.5722980755146517e-06, | |
| "loss": 12.4157, | |
| "step": 177 | |
| }, | |
| { | |
| "epoch": 0.11535968891769281, | |
| "grad_norm": 0.0750642940402031, | |
| "learning_rate": 3.271776770026963e-06, | |
| "loss": 12.4173, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 0.11600777705767984, | |
| "grad_norm": 0.09062527120113373, | |
| "learning_rate": 2.9840304941919415e-06, | |
| "loss": 12.4134, | |
| "step": 179 | |
| }, | |
| { | |
| "epoch": 0.11665586519766688, | |
| "grad_norm": 0.0998246967792511, | |
| "learning_rate": 2.7091379149682685e-06, | |
| "loss": 12.4178, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.11730395333765392, | |
| "grad_norm": 0.09559626132249832, | |
| "learning_rate": 2.4471741852423237e-06, | |
| "loss": 12.4285, | |
| "step": 181 | |
| }, | |
| { | |
| "epoch": 0.11795204147764096, | |
| "grad_norm": 0.08455555140972137, | |
| "learning_rate": 2.1982109232821178e-06, | |
| "loss": 12.4155, | |
| "step": 182 | |
| }, | |
| { | |
| "epoch": 0.118600129617628, | |
| "grad_norm": 0.08278302848339081, | |
| "learning_rate": 1.962316193157593e-06, | |
| "loss": 12.4165, | |
| "step": 183 | |
| }, | |
| { | |
| "epoch": 0.11924821775761503, | |
| "grad_norm": 0.09319792687892914, | |
| "learning_rate": 1.7395544861325718e-06, | |
| "loss": 12.4234, | |
| "step": 184 | |
| }, | |
| { | |
| "epoch": 0.11989630589760207, | |
| "grad_norm": 0.09440692514181137, | |
| "learning_rate": 1.5299867030334814e-06, | |
| "loss": 12.4166, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.12054439403758911, | |
| "grad_norm": 0.08676104247570038, | |
| "learning_rate": 1.333670137599713e-06, | |
| "loss": 12.412, | |
| "step": 186 | |
| }, | |
| { | |
| "epoch": 0.12119248217757615, | |
| "grad_norm": 0.08075446635484695, | |
| "learning_rate": 1.1506584608200367e-06, | |
| "loss": 12.4177, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 0.1218405703175632, | |
| "grad_norm": 0.09502189606428146, | |
| "learning_rate": 9.810017062595322e-07, | |
| "loss": 12.4175, | |
| "step": 188 | |
| }, | |
| { | |
| "epoch": 0.12248865845755022, | |
| "grad_norm": 0.11180531978607178, | |
| "learning_rate": 8.247462563808817e-07, | |
| "loss": 12.4187, | |
| "step": 189 | |
| }, | |
| { | |
| "epoch": 0.12313674659753726, | |
| "grad_norm": 0.08186094462871552, | |
| "learning_rate": 6.819348298638839e-07, | |
| "loss": 12.4251, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.1237848347375243, | |
| "grad_norm": 0.10143885761499405, | |
| "learning_rate": 5.526064699265753e-07, | |
| "loss": 12.4201, | |
| "step": 191 | |
| }, | |
| { | |
| "epoch": 0.12443292287751134, | |
| "grad_norm": 0.09049022942781448, | |
| "learning_rate": 4.367965336512403e-07, | |
| "loss": 12.4148, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 0.12508101101749838, | |
| "grad_norm": 0.07768399268388748, | |
| "learning_rate": 3.3453668231809286e-07, | |
| "loss": 12.4231, | |
| "step": 193 | |
| }, | |
| { | |
| "epoch": 0.12572909915748542, | |
| "grad_norm": 0.09553659707307816, | |
| "learning_rate": 2.458548727494292e-07, | |
| "loss": 12.4256, | |
| "step": 194 | |
| }, | |
| { | |
| "epoch": 0.12637718729747247, | |
| "grad_norm": 0.10629168897867203, | |
| "learning_rate": 1.7077534966650766e-07, | |
| "loss": 12.4154, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.1270252754374595, | |
| "grad_norm": 0.08406480401754379, | |
| "learning_rate": 1.0931863906127327e-07, | |
| "loss": 12.4236, | |
| "step": 196 | |
| }, | |
| { | |
| "epoch": 0.12767336357744652, | |
| "grad_norm": 0.08196214586496353, | |
| "learning_rate": 6.150154258476315e-08, | |
| "loss": 12.4146, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 0.12832145171743356, | |
| "grad_norm": 0.0884324312210083, | |
| "learning_rate": 2.7337132953697554e-08, | |
| "loss": 12.4138, | |
| "step": 198 | |
| }, | |
| { | |
| "epoch": 0.1289695398574206, | |
| "grad_norm": 0.09367944300174713, | |
| "learning_rate": 6.834750376549792e-09, | |
| "loss": 12.4111, | |
| "step": 199 | |
| }, | |
| { | |
| "epoch": 0.12961762799740764, | |
| "grad_norm": 0.09953802078962326, | |
| "learning_rate": 0.0, | |
| "loss": 12.4154, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.12961762799740764, | |
| "eval_loss": 12.417525291442871, | |
| "eval_runtime": 25.653, | |
| "eval_samples_per_second": 101.314, | |
| "eval_steps_per_second": 25.338, | |
| "step": 200 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 965129797632.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |