| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 1370, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18248175182481752, | |
| "grad_norm": 7.047706604003906, | |
| "learning_rate": 0.00044, | |
| "loss": 3.5497, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.36496350364963503, | |
| "grad_norm": 1.9511065483093262, | |
| "learning_rate": 0.00094, | |
| "loss": 0.9894, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5474452554744526, | |
| "grad_norm": 0.2947429418563843, | |
| "learning_rate": 0.0009833333333333332, | |
| "loss": 0.1823, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.7299270072992701, | |
| "grad_norm": 0.8360728025436401, | |
| "learning_rate": 0.0009643939393939394, | |
| "loss": 0.1486, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.9124087591240876, | |
| "grad_norm": 0.44553595781326294, | |
| "learning_rate": 0.0009454545454545454, | |
| "loss": 0.1198, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.1716090887784958, | |
| "eval_runtime": 336.7759, | |
| "eval_samples_per_second": 2.429, | |
| "eval_steps_per_second": 0.407, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 1.094890510948905, | |
| "grad_norm": 0.4483061134815216, | |
| "learning_rate": 0.0009265151515151515, | |
| "loss": 0.1271, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.2773722627737225, | |
| "grad_norm": 0.8258382678031921, | |
| "learning_rate": 0.0009075757575757576, | |
| "loss": 0.0949, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.4598540145985401, | |
| "grad_norm": 0.4852953851222992, | |
| "learning_rate": 0.0008886363636363637, | |
| "loss": 0.0855, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.6423357664233578, | |
| "grad_norm": 0.8936036229133606, | |
| "learning_rate": 0.0008696969696969697, | |
| "loss": 0.0953, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.8248175182481752, | |
| "grad_norm": 0.5471451282501221, | |
| "learning_rate": 0.0008507575757575757, | |
| "loss": 0.0915, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.11422067135572433, | |
| "eval_runtime": 333.7404, | |
| "eval_samples_per_second": 2.451, | |
| "eval_steps_per_second": 0.41, | |
| "step": 274 | |
| }, | |
| { | |
| "epoch": 2.0072992700729926, | |
| "grad_norm": 1.010757565498352, | |
| "learning_rate": 0.0008318181818181818, | |
| "loss": 0.0912, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.18978102189781, | |
| "grad_norm": 1.8606435060501099, | |
| "learning_rate": 0.0008128787878787879, | |
| "loss": 0.0771, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.372262773722628, | |
| "grad_norm": 0.6466093063354492, | |
| "learning_rate": 0.0007939393939393939, | |
| "loss": 0.0768, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.554744525547445, | |
| "grad_norm": 0.4432876408100128, | |
| "learning_rate": 0.0007750000000000001, | |
| "loss": 0.0712, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.7372262773722627, | |
| "grad_norm": 0.5546942949295044, | |
| "learning_rate": 0.0007560606060606061, | |
| "loss": 0.0655, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 2.9197080291970803, | |
| "grad_norm": 0.9572476148605347, | |
| "learning_rate": 0.0007371212121212121, | |
| "loss": 0.0769, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.044818274676799774, | |
| "eval_runtime": 337.2883, | |
| "eval_samples_per_second": 2.425, | |
| "eval_steps_per_second": 0.406, | |
| "step": 411 | |
| }, | |
| { | |
| "epoch": 3.102189781021898, | |
| "grad_norm": 0.23699763417243958, | |
| "learning_rate": 0.0007181818181818181, | |
| "loss": 0.0416, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 3.2846715328467155, | |
| "grad_norm": 0.2720116078853607, | |
| "learning_rate": 0.0006992424242424243, | |
| "loss": 0.043, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.4671532846715327, | |
| "grad_norm": 0.1873721480369568, | |
| "learning_rate": 0.0006803030303030303, | |
| "loss": 0.0344, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.6496350364963503, | |
| "grad_norm": 0.3385811150074005, | |
| "learning_rate": 0.0006613636363636364, | |
| "loss": 0.0455, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.832116788321168, | |
| "grad_norm": 0.27543193101882935, | |
| "learning_rate": 0.0006424242424242425, | |
| "loss": 0.0444, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.034519024193286896, | |
| "eval_runtime": 335.9, | |
| "eval_samples_per_second": 2.435, | |
| "eval_steps_per_second": 0.408, | |
| "step": 548 | |
| }, | |
| { | |
| "epoch": 4.014598540145985, | |
| "grad_norm": 0.08995059132575989, | |
| "learning_rate": 0.0006234848484848486, | |
| "loss": 0.0502, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 4.197080291970803, | |
| "grad_norm": 0.4935706853866577, | |
| "learning_rate": 0.0006045454545454545, | |
| "loss": 0.0378, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 4.37956204379562, | |
| "grad_norm": 0.47609230875968933, | |
| "learning_rate": 0.0005856060606060606, | |
| "loss": 0.0262, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.562043795620438, | |
| "grad_norm": 0.5216805934906006, | |
| "learning_rate": 0.0005666666666666667, | |
| "loss": 0.036, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 4.744525547445256, | |
| "grad_norm": 0.2257470041513443, | |
| "learning_rate": 0.0005477272727272728, | |
| "loss": 0.0264, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 4.927007299270073, | |
| "grad_norm": 0.5467531085014343, | |
| "learning_rate": 0.0005287878787878788, | |
| "loss": 0.0439, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.02093869261443615, | |
| "eval_runtime": 338.3118, | |
| "eval_samples_per_second": 2.418, | |
| "eval_steps_per_second": 0.405, | |
| "step": 685 | |
| }, | |
| { | |
| "epoch": 5.109489051094891, | |
| "grad_norm": 0.18127380311489105, | |
| "learning_rate": 0.0005098484848484848, | |
| "loss": 0.0205, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.291970802919708, | |
| "grad_norm": 0.10721493512392044, | |
| "learning_rate": 0.0004909090909090909, | |
| "loss": 0.0298, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 5.474452554744525, | |
| "grad_norm": 0.1958245486021042, | |
| "learning_rate": 0.000471969696969697, | |
| "loss": 0.0226, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 5.656934306569343, | |
| "grad_norm": 0.04980677366256714, | |
| "learning_rate": 0.000453030303030303, | |
| "loss": 0.0188, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 5.839416058394161, | |
| "grad_norm": 0.3205123841762543, | |
| "learning_rate": 0.0004340909090909091, | |
| "loss": 0.0175, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.011000733822584152, | |
| "eval_runtime": 338.6304, | |
| "eval_samples_per_second": 2.416, | |
| "eval_steps_per_second": 0.405, | |
| "step": 822 | |
| }, | |
| { | |
| "epoch": 6.021897810218978, | |
| "grad_norm": 0.01701493188738823, | |
| "learning_rate": 0.0004151515151515152, | |
| "loss": 0.0183, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 6.204379562043796, | |
| "grad_norm": 0.07489558309316635, | |
| "learning_rate": 0.00039621212121212124, | |
| "loss": 0.0122, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 6.386861313868613, | |
| "grad_norm": 0.23450659215450287, | |
| "learning_rate": 0.0003772727272727273, | |
| "loss": 0.0132, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 6.569343065693431, | |
| "grad_norm": 0.14261268079280853, | |
| "learning_rate": 0.00035833333333333333, | |
| "loss": 0.0109, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 6.751824817518248, | |
| "grad_norm": 0.2159743756055832, | |
| "learning_rate": 0.00033939393939393943, | |
| "loss": 0.0118, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 6.934306569343065, | |
| "grad_norm": 0.3018836975097656, | |
| "learning_rate": 0.00032045454545454543, | |
| "loss": 0.0136, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.007896244525909424, | |
| "eval_runtime": 336.2346, | |
| "eval_samples_per_second": 2.433, | |
| "eval_steps_per_second": 0.407, | |
| "step": 959 | |
| }, | |
| { | |
| "epoch": 7.116788321167883, | |
| "grad_norm": 0.010803497396409512, | |
| "learning_rate": 0.00030151515151515153, | |
| "loss": 0.0055, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 7.299270072992701, | |
| "grad_norm": 0.09972184151411057, | |
| "learning_rate": 0.0002825757575757576, | |
| "loss": 0.0077, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 7.481751824817518, | |
| "grad_norm": 0.17541509866714478, | |
| "learning_rate": 0.0002636363636363636, | |
| "loss": 0.0066, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 7.664233576642336, | |
| "grad_norm": 0.025586027652025223, | |
| "learning_rate": 0.00024469696969696967, | |
| "loss": 0.0115, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 7.846715328467154, | |
| "grad_norm": 0.005738795734941959, | |
| "learning_rate": 0.00022575757575757577, | |
| "loss": 0.0067, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.003122333437204361, | |
| "eval_runtime": 348.1783, | |
| "eval_samples_per_second": 2.349, | |
| "eval_steps_per_second": 0.393, | |
| "step": 1096 | |
| }, | |
| { | |
| "epoch": 8.02919708029197, | |
| "grad_norm": 0.0272404532879591, | |
| "learning_rate": 0.00020681818181818182, | |
| "loss": 0.0042, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 8.211678832116789, | |
| "grad_norm": 0.28844815492630005, | |
| "learning_rate": 0.00018787878787878787, | |
| "loss": 0.0048, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 8.394160583941606, | |
| "grad_norm": 0.009899291209876537, | |
| "learning_rate": 0.00016893939393939394, | |
| "loss": 0.0055, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 8.576642335766424, | |
| "grad_norm": 0.06182921677827835, | |
| "learning_rate": 0.00015, | |
| "loss": 0.0031, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 8.75912408759124, | |
| "grad_norm": 0.009944899007678032, | |
| "learning_rate": 0.0001310606060606061, | |
| "loss": 0.0027, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 8.941605839416058, | |
| "grad_norm": 0.020729435607790947, | |
| "learning_rate": 0.00011212121212121212, | |
| "loss": 0.0024, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.0016915486194193363, | |
| "eval_runtime": 338.3118, | |
| "eval_samples_per_second": 2.418, | |
| "eval_steps_per_second": 0.405, | |
| "step": 1233 | |
| }, | |
| { | |
| "epoch": 9.124087591240876, | |
| "grad_norm": 0.059274736791849136, | |
| "learning_rate": 9.318181818181818e-05, | |
| "loss": 0.0016, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 9.306569343065693, | |
| "grad_norm": 0.023554878309369087, | |
| "learning_rate": 7.424242424242426e-05, | |
| "loss": 0.002, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 9.489051094890511, | |
| "grad_norm": 0.01931552216410637, | |
| "learning_rate": 5.5303030303030304e-05, | |
| "loss": 0.0009, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 9.671532846715328, | |
| "grad_norm": 0.0068663558922708035, | |
| "learning_rate": 3.6363636363636364e-05, | |
| "loss": 0.0028, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 9.854014598540147, | |
| "grad_norm": 0.014311350882053375, | |
| "learning_rate": 1.7424242424242425e-05, | |
| "loss": 0.0015, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.0014132640790194273, | |
| "eval_runtime": 338.1601, | |
| "eval_samples_per_second": 2.419, | |
| "eval_steps_per_second": 0.405, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 1370, | |
| "total_flos": 8.4597106212864e+18, | |
| "train_loss": 0.12102186830755132, | |
| "train_runtime": 7917.0938, | |
| "train_samples_per_second": 1.033, | |
| "train_steps_per_second": 0.173 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1370, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8.4597106212864e+18, | |
| "train_batch_size": 6, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |