| { | |
| "best_metric": 2.7241549491882324, | |
| "best_model_checkpoint": "./results/models/checkpoint-26406", | |
| "epoch": 27.0, | |
| "eval_steps": 500, | |
| "global_step": 26406, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.5112474437627812, | |
| "grad_norm": 0.1318359375, | |
| "learning_rate": 0.0009897750511247444, | |
| "loss": 3.8641, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 3.6347575187683105, | |
| "eval_runtime": 0.3183, | |
| "eval_samples_per_second": 1571.027, | |
| "eval_steps_per_second": 3.142, | |
| "step": 978 | |
| }, | |
| { | |
| "epoch": 1.0224948875255624, | |
| "grad_norm": 0.2255859375, | |
| "learning_rate": 0.0009795501022494889, | |
| "loss": 3.6938, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.5337423312883436, | |
| "grad_norm": 0.263671875, | |
| "learning_rate": 0.0009693251533742331, | |
| "loss": 3.488, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 3.214155912399292, | |
| "eval_runtime": 0.95, | |
| "eval_samples_per_second": 526.301, | |
| "eval_steps_per_second": 1.053, | |
| "step": 1956 | |
| }, | |
| { | |
| "epoch": 2.044989775051125, | |
| "grad_norm": 0.224609375, | |
| "learning_rate": 0.0009591002044989775, | |
| "loss": 3.2879, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.556237218813906, | |
| "grad_norm": 0.2373046875, | |
| "learning_rate": 0.0009488752556237219, | |
| "loss": 3.1641, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 3.0714735984802246, | |
| "eval_runtime": 0.9223, | |
| "eval_samples_per_second": 542.135, | |
| "eval_steps_per_second": 1.084, | |
| "step": 2934 | |
| }, | |
| { | |
| "epoch": 3.067484662576687, | |
| "grad_norm": 0.29296875, | |
| "learning_rate": 0.0009386503067484663, | |
| "loss": 3.0912, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.5787321063394684, | |
| "grad_norm": 0.255859375, | |
| "learning_rate": 0.0009284253578732107, | |
| "loss": 3.0323, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.990953207015991, | |
| "eval_runtime": 1.0538, | |
| "eval_samples_per_second": 474.475, | |
| "eval_steps_per_second": 0.949, | |
| "step": 3912 | |
| }, | |
| { | |
| "epoch": 4.08997955010225, | |
| "grad_norm": 0.259765625, | |
| "learning_rate": 0.000918200408997955, | |
| "loss": 2.9912, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.601226993865031, | |
| "grad_norm": 0.337890625, | |
| "learning_rate": 0.0009079754601226994, | |
| "loss": 2.9553, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 2.9412269592285156, | |
| "eval_runtime": 1.0599, | |
| "eval_samples_per_second": 471.73, | |
| "eval_steps_per_second": 0.943, | |
| "step": 4890 | |
| }, | |
| { | |
| "epoch": 5.112474437627812, | |
| "grad_norm": 0.26953125, | |
| "learning_rate": 0.0008977505112474438, | |
| "loss": 2.9255, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.623721881390593, | |
| "grad_norm": 0.25390625, | |
| "learning_rate": 0.0008875255623721882, | |
| "loss": 2.8991, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 2.905097723007202, | |
| "eval_runtime": 1.0476, | |
| "eval_samples_per_second": 477.298, | |
| "eval_steps_per_second": 0.955, | |
| "step": 5868 | |
| }, | |
| { | |
| "epoch": 6.134969325153374, | |
| "grad_norm": 0.3203125, | |
| "learning_rate": 0.0008773006134969325, | |
| "loss": 2.8794, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.6462167689161555, | |
| "grad_norm": 0.29296875, | |
| "learning_rate": 0.0008670756646216768, | |
| "loss": 2.8577, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 2.8703627586364746, | |
| "eval_runtime": 1.0515, | |
| "eval_samples_per_second": 475.508, | |
| "eval_steps_per_second": 0.951, | |
| "step": 6846 | |
| }, | |
| { | |
| "epoch": 7.157464212678937, | |
| "grad_norm": 0.29296875, | |
| "learning_rate": 0.0008568507157464213, | |
| "loss": 2.8435, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.668711656441718, | |
| "grad_norm": 0.30078125, | |
| "learning_rate": 0.0008466257668711657, | |
| "loss": 2.8289, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 2.847052812576294, | |
| "eval_runtime": 1.0464, | |
| "eval_samples_per_second": 477.826, | |
| "eval_steps_per_second": 0.956, | |
| "step": 7824 | |
| }, | |
| { | |
| "epoch": 8.1799591002045, | |
| "grad_norm": 0.3046875, | |
| "learning_rate": 0.0008364008179959101, | |
| "loss": 2.8131, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.69120654396728, | |
| "grad_norm": 0.359375, | |
| "learning_rate": 0.0008261758691206544, | |
| "loss": 2.8022, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 2.8293135166168213, | |
| "eval_runtime": 0.9764, | |
| "eval_samples_per_second": 512.074, | |
| "eval_steps_per_second": 1.024, | |
| "step": 8802 | |
| }, | |
| { | |
| "epoch": 9.202453987730062, | |
| "grad_norm": 0.310546875, | |
| "learning_rate": 0.0008159509202453987, | |
| "loss": 2.7914, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.713701431492842, | |
| "grad_norm": 0.314453125, | |
| "learning_rate": 0.0008057259713701431, | |
| "loss": 2.7813, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 2.8122880458831787, | |
| "eval_runtime": 1.0064, | |
| "eval_samples_per_second": 496.84, | |
| "eval_steps_per_second": 0.994, | |
| "step": 9780 | |
| }, | |
| { | |
| "epoch": 10.224948875255624, | |
| "grad_norm": 0.296875, | |
| "learning_rate": 0.0007955010224948876, | |
| "loss": 2.7682, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.736196319018404, | |
| "grad_norm": 0.328125, | |
| "learning_rate": 0.0007852760736196319, | |
| "loss": 2.7607, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 2.8023903369903564, | |
| "eval_runtime": 1.0601, | |
| "eval_samples_per_second": 471.642, | |
| "eval_steps_per_second": 0.943, | |
| "step": 10758 | |
| }, | |
| { | |
| "epoch": 11.247443762781186, | |
| "grad_norm": 0.294921875, | |
| "learning_rate": 0.0007750511247443763, | |
| "loss": 2.7527, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 11.758691206543967, | |
| "grad_norm": 0.3046875, | |
| "learning_rate": 0.0007648261758691206, | |
| "loss": 2.7443, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 2.7869932651519775, | |
| "eval_runtime": 1.0222, | |
| "eval_samples_per_second": 489.149, | |
| "eval_steps_per_second": 0.978, | |
| "step": 11736 | |
| }, | |
| { | |
| "epoch": 12.269938650306749, | |
| "grad_norm": 0.298828125, | |
| "learning_rate": 0.000754601226993865, | |
| "loss": 2.7366, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 12.781186094069529, | |
| "grad_norm": 0.34375, | |
| "learning_rate": 0.0007443762781186095, | |
| "loss": 2.7313, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 2.785067081451416, | |
| "eval_runtime": 1.051, | |
| "eval_samples_per_second": 475.716, | |
| "eval_steps_per_second": 0.951, | |
| "step": 12714 | |
| }, | |
| { | |
| "epoch": 13.292433537832311, | |
| "grad_norm": 0.33203125, | |
| "learning_rate": 0.0007341513292433538, | |
| "loss": 2.7223, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 13.803680981595091, | |
| "grad_norm": 0.33203125, | |
| "learning_rate": 0.0007239263803680982, | |
| "loss": 2.7179, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 2.772521495819092, | |
| "eval_runtime": 1.0385, | |
| "eval_samples_per_second": 481.485, | |
| "eval_steps_per_second": 0.963, | |
| "step": 13692 | |
| }, | |
| { | |
| "epoch": 14.314928425357873, | |
| "grad_norm": 0.330078125, | |
| "learning_rate": 0.0007137014314928425, | |
| "loss": 2.7088, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 14.826175869120654, | |
| "grad_norm": 0.34375, | |
| "learning_rate": 0.0007034764826175869, | |
| "loss": 2.7051, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_loss": 2.7638583183288574, | |
| "eval_runtime": 0.9397, | |
| "eval_samples_per_second": 532.097, | |
| "eval_steps_per_second": 1.064, | |
| "step": 14670 | |
| }, | |
| { | |
| "epoch": 15.337423312883436, | |
| "grad_norm": 0.3203125, | |
| "learning_rate": 0.0006932515337423313, | |
| "loss": 2.6999, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 15.848670756646216, | |
| "grad_norm": 0.31640625, | |
| "learning_rate": 0.0006830265848670757, | |
| "loss": 2.6961, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 2.7651634216308594, | |
| "eval_runtime": 0.9625, | |
| "eval_samples_per_second": 519.471, | |
| "eval_steps_per_second": 1.039, | |
| "step": 15648 | |
| }, | |
| { | |
| "epoch": 16.359918200409, | |
| "grad_norm": 0.388671875, | |
| "learning_rate": 0.0006728016359918201, | |
| "loss": 2.6903, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 16.87116564417178, | |
| "grad_norm": 0.345703125, | |
| "learning_rate": 0.0006625766871165644, | |
| "loss": 2.6905, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_loss": 2.75215744972229, | |
| "eval_runtime": 0.8724, | |
| "eval_samples_per_second": 573.163, | |
| "eval_steps_per_second": 1.146, | |
| "step": 16626 | |
| }, | |
| { | |
| "epoch": 17.38241308793456, | |
| "grad_norm": 0.337890625, | |
| "learning_rate": 0.0006523517382413088, | |
| "loss": 2.6791, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 17.893660531697343, | |
| "grad_norm": 0.328125, | |
| "learning_rate": 0.0006421267893660531, | |
| "loss": 2.6817, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_loss": 2.750427722930908, | |
| "eval_runtime": 1.0601, | |
| "eval_samples_per_second": 471.655, | |
| "eval_steps_per_second": 0.943, | |
| "step": 17604 | |
| }, | |
| { | |
| "epoch": 18.404907975460123, | |
| "grad_norm": 0.33984375, | |
| "learning_rate": 0.0006319018404907975, | |
| "loss": 2.6712, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 18.916155419222903, | |
| "grad_norm": 0.349609375, | |
| "learning_rate": 0.000621676891615542, | |
| "loss": 2.6721, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_loss": 2.740670919418335, | |
| "eval_runtime": 0.9797, | |
| "eval_samples_per_second": 510.381, | |
| "eval_steps_per_second": 1.021, | |
| "step": 18582 | |
| }, | |
| { | |
| "epoch": 19.427402862985684, | |
| "grad_norm": 0.328125, | |
| "learning_rate": 0.0006114519427402863, | |
| "loss": 2.6632, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 19.938650306748468, | |
| "grad_norm": 0.36328125, | |
| "learning_rate": 0.0006012269938650307, | |
| "loss": 2.6677, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_loss": 2.740766763687134, | |
| "eval_runtime": 0.9358, | |
| "eval_samples_per_second": 534.326, | |
| "eval_steps_per_second": 1.069, | |
| "step": 19560 | |
| }, | |
| { | |
| "epoch": 20.449897750511248, | |
| "grad_norm": 0.345703125, | |
| "learning_rate": 0.000591002044989775, | |
| "loss": 2.6576, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 20.961145194274028, | |
| "grad_norm": 0.349609375, | |
| "learning_rate": 0.0005807770961145194, | |
| "loss": 2.6624, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_loss": 2.7370216846466064, | |
| "eval_runtime": 0.9579, | |
| "eval_samples_per_second": 521.987, | |
| "eval_steps_per_second": 1.044, | |
| "step": 20538 | |
| }, | |
| { | |
| "epoch": 21.47239263803681, | |
| "grad_norm": 0.314453125, | |
| "learning_rate": 0.0005705521472392639, | |
| "loss": 2.65, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 21.983640081799592, | |
| "grad_norm": 0.341796875, | |
| "learning_rate": 0.0005603271983640082, | |
| "loss": 2.6575, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_loss": 2.734588384628296, | |
| "eval_runtime": 0.8297, | |
| "eval_samples_per_second": 602.609, | |
| "eval_steps_per_second": 1.205, | |
| "step": 21516 | |
| }, | |
| { | |
| "epoch": 22.494887525562373, | |
| "grad_norm": 0.349609375, | |
| "learning_rate": 0.0005501022494887526, | |
| "loss": 2.6453, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_loss": 2.73508882522583, | |
| "eval_runtime": 1.0104, | |
| "eval_samples_per_second": 494.868, | |
| "eval_steps_per_second": 0.99, | |
| "step": 22494 | |
| }, | |
| { | |
| "epoch": 23.006134969325153, | |
| "grad_norm": 0.345703125, | |
| "learning_rate": 0.0005398773006134969, | |
| "loss": 2.6533, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 23.517382413087933, | |
| "grad_norm": 0.33984375, | |
| "learning_rate": 0.0005296523517382413, | |
| "loss": 2.6441, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_loss": 2.731626033782959, | |
| "eval_runtime": 0.674, | |
| "eval_samples_per_second": 741.886, | |
| "eval_steps_per_second": 1.484, | |
| "step": 23472 | |
| }, | |
| { | |
| "epoch": 24.028629856850717, | |
| "grad_norm": 0.322265625, | |
| "learning_rate": 0.0005194274028629857, | |
| "loss": 2.6442, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 24.539877300613497, | |
| "grad_norm": 0.330078125, | |
| "learning_rate": 0.00050920245398773, | |
| "loss": 2.637, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_loss": 2.731029510498047, | |
| "eval_runtime": 0.7022, | |
| "eval_samples_per_second": 712.014, | |
| "eval_steps_per_second": 1.424, | |
| "step": 24450 | |
| }, | |
| { | |
| "epoch": 25.051124744376278, | |
| "grad_norm": 0.326171875, | |
| "learning_rate": 0.0004989775051124745, | |
| "loss": 2.6437, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 25.562372188139058, | |
| "grad_norm": 0.333984375, | |
| "learning_rate": 0.0004887525562372188, | |
| "loss": 2.6364, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_loss": 2.7300286293029785, | |
| "eval_runtime": 0.8078, | |
| "eval_samples_per_second": 618.965, | |
| "eval_steps_per_second": 1.238, | |
| "step": 25428 | |
| }, | |
| { | |
| "epoch": 26.073619631901842, | |
| "grad_norm": 0.357421875, | |
| "learning_rate": 0.0004785276073619632, | |
| "loss": 2.6382, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 26.584867075664622, | |
| "grad_norm": 0.330078125, | |
| "learning_rate": 0.0004683026584867076, | |
| "loss": 2.632, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_loss": 2.7241549491882324, | |
| "eval_runtime": 0.8157, | |
| "eval_samples_per_second": 612.964, | |
| "eval_steps_per_second": 1.226, | |
| "step": 26406 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 48900, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "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": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.553177237041664e+17, | |
| "train_batch_size": 1024, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |