| { | |
| "best_metric": 0.4678440617843855, | |
| "best_model_checkpoint": "/fsx/anton/cosmopedia/edu_score/bert_snowflake_regression_4/checkpoint-7000", | |
| "epoch": 4.247572815533981, | |
| "eval_steps": 1000, | |
| "global_step": 7000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06067961165048544, | |
| "grad_norm": 0.5638211965560913, | |
| "learning_rate": 0.0002990898058252427, | |
| "loss": 0.4753, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12135922330097088, | |
| "grad_norm": 0.47830212116241455, | |
| "learning_rate": 0.0002981796116504854, | |
| "loss": 0.357, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1820388349514563, | |
| "grad_norm": 0.6941384077072144, | |
| "learning_rate": 0.0002972694174757281, | |
| "loss": 0.3542, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.24271844660194175, | |
| "grad_norm": 0.459163635969162, | |
| "learning_rate": 0.00029635922330097087, | |
| "loss": 0.3508, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.30339805825242716, | |
| "grad_norm": 1.1585971117019653, | |
| "learning_rate": 0.0002954490291262136, | |
| "loss": 0.3407, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.3640776699029126, | |
| "grad_norm": 0.6505594849586487, | |
| "learning_rate": 0.0002945388349514563, | |
| "loss": 0.3394, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.42475728155339804, | |
| "grad_norm": 0.9804072976112366, | |
| "learning_rate": 0.000293628640776699, | |
| "loss": 0.3435, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.4854368932038835, | |
| "grad_norm": 0.5816351175308228, | |
| "learning_rate": 0.0002927184466019417, | |
| "loss": 0.3323, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.5461165048543689, | |
| "grad_norm": 0.6582027673721313, | |
| "learning_rate": 0.00029180825242718447, | |
| "loss": 0.3293, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.6067961165048543, | |
| "grad_norm": 0.8432559370994568, | |
| "learning_rate": 0.0002908980582524271, | |
| "loss": 0.3337, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6067961165048543, | |
| "eval_accuracy": 0.6516312117268014, | |
| "eval_f1_macro": 0.388617225018415, | |
| "eval_loss": 0.32428401708602905, | |
| "eval_precision": 0.5273989146868351, | |
| "eval_recall": 0.3731209274235363, | |
| "eval_runtime": 63.5791, | |
| "eval_samples_per_second": 737.144, | |
| "eval_steps_per_second": 5.772, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6674757281553398, | |
| "grad_norm": 0.6387248039245605, | |
| "learning_rate": 0.0002899878640776699, | |
| "loss": 0.3292, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.7281553398058253, | |
| "grad_norm": 2.0111730098724365, | |
| "learning_rate": 0.0002890776699029126, | |
| "loss": 0.3283, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.7888349514563107, | |
| "grad_norm": 1.0319699048995972, | |
| "learning_rate": 0.0002881674757281553, | |
| "loss": 0.3236, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.8495145631067961, | |
| "grad_norm": 1.194286584854126, | |
| "learning_rate": 0.000287257281553398, | |
| "loss": 0.3178, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.9101941747572816, | |
| "grad_norm": 0.4329046308994293, | |
| "learning_rate": 0.00028634708737864073, | |
| "loss": 0.3234, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.970873786407767, | |
| "grad_norm": 0.4490291476249695, | |
| "learning_rate": 0.0002854368932038835, | |
| "loss": 0.3148, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.0315533980582525, | |
| "grad_norm": 1.9341398477554321, | |
| "learning_rate": 0.0002845266990291262, | |
| "loss": 0.3157, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.0922330097087378, | |
| "grad_norm": 0.6705629825592041, | |
| "learning_rate": 0.0002836165048543689, | |
| "loss": 0.3144, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.1529126213592233, | |
| "grad_norm": 0.3708420395851135, | |
| "learning_rate": 0.0002827063106796116, | |
| "loss": 0.3106, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.2135922330097086, | |
| "grad_norm": 0.4166070222854614, | |
| "learning_rate": 0.00028179611650485433, | |
| "loss": 0.3065, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.2135922330097086, | |
| "eval_accuracy": 0.6756993193505025, | |
| "eval_f1_macro": 0.4177625223934091, | |
| "eval_loss": 0.3020932972431183, | |
| "eval_precision": 0.5208365532058327, | |
| "eval_recall": 0.3908095186846785, | |
| "eval_runtime": 63.8686, | |
| "eval_samples_per_second": 733.803, | |
| "eval_steps_per_second": 5.746, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.2742718446601942, | |
| "grad_norm": 0.949810266494751, | |
| "learning_rate": 0.00028088592233009704, | |
| "loss": 0.309, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.3349514563106797, | |
| "grad_norm": 0.6933236718177795, | |
| "learning_rate": 0.00027997572815533975, | |
| "loss": 0.3039, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.395631067961165, | |
| "grad_norm": 0.4874693751335144, | |
| "learning_rate": 0.0002790655339805825, | |
| "loss": 0.3016, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.4563106796116505, | |
| "grad_norm": 0.5307803750038147, | |
| "learning_rate": 0.0002781553398058252, | |
| "loss": 0.295, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.516990291262136, | |
| "grad_norm": 0.7260825634002686, | |
| "learning_rate": 0.00027724514563106793, | |
| "loss": 0.298, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.5776699029126213, | |
| "grad_norm": 1.3546072244644165, | |
| "learning_rate": 0.00027633495145631064, | |
| "loss": 0.2937, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.6383495145631068, | |
| "grad_norm": 0.7695233821868896, | |
| "learning_rate": 0.00027542475728155335, | |
| "loss": 0.2939, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 1.6990291262135924, | |
| "grad_norm": 0.46857160329818726, | |
| "learning_rate": 0.0002745145631067961, | |
| "loss": 0.2911, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.7597087378640777, | |
| "grad_norm": 0.521542489528656, | |
| "learning_rate": 0.0002736043689320388, | |
| "loss": 0.2936, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.820388349514563, | |
| "grad_norm": 1.1797749996185303, | |
| "learning_rate": 0.00027269417475728154, | |
| "loss": 0.291, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.820388349514563, | |
| "eval_accuracy": 0.6800947361683061, | |
| "eval_f1_macro": 0.43005056999542096, | |
| "eval_loss": 0.2904761731624603, | |
| "eval_precision": 0.5283372179121356, | |
| "eval_recall": 0.40012210038254903, | |
| "eval_runtime": 63.5697, | |
| "eval_samples_per_second": 737.253, | |
| "eval_steps_per_second": 5.773, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.8810679611650487, | |
| "grad_norm": 0.9966709613800049, | |
| "learning_rate": 0.00027178398058252425, | |
| "loss": 0.2965, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.941747572815534, | |
| "grad_norm": 0.40996024012565613, | |
| "learning_rate": 0.00027087378640776696, | |
| "loss": 0.2941, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 2.0024271844660193, | |
| "grad_norm": 0.5450060367584229, | |
| "learning_rate": 0.00026996359223300967, | |
| "loss": 0.2912, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 2.063106796116505, | |
| "grad_norm": 0.5307539701461792, | |
| "learning_rate": 0.0002690533980582524, | |
| "loss": 0.2872, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 2.1237864077669903, | |
| "grad_norm": 0.5863193273544312, | |
| "learning_rate": 0.00026814320388349514, | |
| "loss": 0.2929, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.1844660194174756, | |
| "grad_norm": 0.584078311920166, | |
| "learning_rate": 0.00026723300970873785, | |
| "loss": 0.2879, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 2.2451456310679614, | |
| "grad_norm": 0.6381602883338928, | |
| "learning_rate": 0.00026632281553398056, | |
| "loss": 0.2892, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 2.3058252427184467, | |
| "grad_norm": 0.4760149121284485, | |
| "learning_rate": 0.00026541262135922327, | |
| "loss": 0.2863, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 2.366504854368932, | |
| "grad_norm": 0.4088296890258789, | |
| "learning_rate": 0.000264502427184466, | |
| "loss": 0.2913, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 2.4271844660194173, | |
| "grad_norm": 1.3476176261901855, | |
| "learning_rate": 0.00026359223300970874, | |
| "loss": 0.2845, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.4271844660194173, | |
| "eval_accuracy": 0.6929182580493738, | |
| "eval_f1_macro": 0.4442664055274559, | |
| "eval_loss": 0.280377060174942, | |
| "eval_precision": 0.5328225535144124, | |
| "eval_recall": 0.41160029395774395, | |
| "eval_runtime": 64.199, | |
| "eval_samples_per_second": 730.027, | |
| "eval_steps_per_second": 5.717, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.487864077669903, | |
| "grad_norm": 0.6318752765655518, | |
| "learning_rate": 0.0002626820388349514, | |
| "loss": 0.2812, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 2.5485436893203883, | |
| "grad_norm": 0.49435973167419434, | |
| "learning_rate": 0.00026177184466019416, | |
| "loss": 0.2803, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 2.6092233009708736, | |
| "grad_norm": 0.4300900101661682, | |
| "learning_rate": 0.00026086165048543687, | |
| "loss": 0.2853, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 2.6699029126213594, | |
| "grad_norm": 0.9545436501502991, | |
| "learning_rate": 0.0002599514563106796, | |
| "loss": 0.2813, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 2.7305825242718447, | |
| "grad_norm": 0.5803716778755188, | |
| "learning_rate": 0.0002590412621359223, | |
| "loss": 0.2838, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.79126213592233, | |
| "grad_norm": 1.4714713096618652, | |
| "learning_rate": 0.000258131067961165, | |
| "loss": 0.2814, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 2.8519417475728153, | |
| "grad_norm": 0.6767821311950684, | |
| "learning_rate": 0.00025722087378640777, | |
| "loss": 0.2741, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 2.912621359223301, | |
| "grad_norm": 0.4653462767601013, | |
| "learning_rate": 0.0002563106796116505, | |
| "loss": 0.2783, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.9733009708737863, | |
| "grad_norm": 1.3012775182724, | |
| "learning_rate": 0.0002554004854368932, | |
| "loss": 0.283, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 3.033980582524272, | |
| "grad_norm": 0.4733451306819916, | |
| "learning_rate": 0.0002544902912621359, | |
| "loss": 0.2767, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.033980582524272, | |
| "eval_accuracy": 0.6949666076343696, | |
| "eval_f1_macro": 0.4524166681181929, | |
| "eval_loss": 0.2772601842880249, | |
| "eval_precision": 0.52914110464024, | |
| "eval_recall": 0.42261704559523156, | |
| "eval_runtime": 63.4257, | |
| "eval_samples_per_second": 738.927, | |
| "eval_steps_per_second": 5.786, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.0946601941747574, | |
| "grad_norm": 0.4103662371635437, | |
| "learning_rate": 0.0002535800970873786, | |
| "loss": 0.2796, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 3.1553398058252426, | |
| "grad_norm": 0.4195462763309479, | |
| "learning_rate": 0.0002526699029126213, | |
| "loss": 0.2757, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 3.216019417475728, | |
| "grad_norm": 1.2391552925109863, | |
| "learning_rate": 0.0002517597087378641, | |
| "loss": 0.2783, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 3.2766990291262137, | |
| "grad_norm": 1.2029412984848022, | |
| "learning_rate": 0.0002508495145631068, | |
| "loss": 0.2772, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 3.337378640776699, | |
| "grad_norm": 0.5050978660583496, | |
| "learning_rate": 0.0002499393203883495, | |
| "loss": 0.2776, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.3980582524271843, | |
| "grad_norm": 1.0107412338256836, | |
| "learning_rate": 0.0002490291262135922, | |
| "loss": 0.2766, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 3.45873786407767, | |
| "grad_norm": 0.4374917149543762, | |
| "learning_rate": 0.0002481189320388349, | |
| "loss": 0.2719, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 3.5194174757281553, | |
| "grad_norm": 1.6768765449523926, | |
| "learning_rate": 0.0002472087378640777, | |
| "loss": 0.2803, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 3.5800970873786406, | |
| "grad_norm": 0.8120823502540588, | |
| "learning_rate": 0.0002462985436893204, | |
| "loss": 0.2723, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 3.6407766990291264, | |
| "grad_norm": 1.3967177867889404, | |
| "learning_rate": 0.0002453883495145631, | |
| "loss": 0.2796, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.6407766990291264, | |
| "eval_accuracy": 0.7001301555465466, | |
| "eval_f1_macro": 0.46180498265852127, | |
| "eval_loss": 0.272257536649704, | |
| "eval_precision": 0.5281578618931982, | |
| "eval_recall": 0.4315295129889904, | |
| "eval_runtime": 63.8213, | |
| "eval_samples_per_second": 734.347, | |
| "eval_steps_per_second": 5.75, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.7014563106796117, | |
| "grad_norm": 0.6093985438346863, | |
| "learning_rate": 0.0002444781553398058, | |
| "loss": 0.2744, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 3.762135922330097, | |
| "grad_norm": 0.7282202243804932, | |
| "learning_rate": 0.00024356796116504852, | |
| "loss": 0.2715, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 3.8228155339805827, | |
| "grad_norm": 1.1341967582702637, | |
| "learning_rate": 0.00024265776699029123, | |
| "loss": 0.2709, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 3.883495145631068, | |
| "grad_norm": 0.8576841354370117, | |
| "learning_rate": 0.00024174757281553394, | |
| "loss": 0.275, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 3.9441747572815533, | |
| "grad_norm": 0.5656840205192566, | |
| "learning_rate": 0.00024083737864077668, | |
| "loss": 0.2676, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 4.004854368932039, | |
| "grad_norm": 0.6544743180274963, | |
| "learning_rate": 0.0002399271844660194, | |
| "loss": 0.2734, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 4.065533980582524, | |
| "grad_norm": 1.5159205198287964, | |
| "learning_rate": 0.0002390169902912621, | |
| "loss": 0.2666, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 4.12621359223301, | |
| "grad_norm": 0.9112799763679504, | |
| "learning_rate": 0.00023810679611650483, | |
| "loss": 0.2646, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 4.186893203883495, | |
| "grad_norm": 0.6971092224121094, | |
| "learning_rate": 0.00023719660194174754, | |
| "loss": 0.2681, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 4.247572815533981, | |
| "grad_norm": 0.5126680731773376, | |
| "learning_rate": 0.00023628640776699028, | |
| "loss": 0.2669, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 4.247572815533981, | |
| "eval_accuracy": 0.7024985597541981, | |
| "eval_f1_macro": 0.4678440617843855, | |
| "eval_loss": 0.26851192116737366, | |
| "eval_precision": 0.5330566032345198, | |
| "eval_recall": 0.43620949446199525, | |
| "eval_runtime": 63.9009, | |
| "eval_samples_per_second": 733.432, | |
| "eval_steps_per_second": 5.743, | |
| "step": 7000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 32960, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 1000, | |
| "total_flos": 4.713992160165028e+17, | |
| "train_batch_size": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |