| { | |
| "best_metric": 0.8985657508208054, | |
| "best_model_checkpoint": "CHECKPOINTS/checkpoint-3307", | |
| "epoch": 11.083056143533918, | |
| "eval_steps": 3307, | |
| "global_step": 19842, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02519907267412559, | |
| "grad_norm": 0.7393302321434021, | |
| "learning_rate": 2.4899193548387098e-06, | |
| "loss": 0.3059, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.05039814534825118, | |
| "grad_norm": 0.6863130331039429, | |
| "learning_rate": 5.010080645161291e-06, | |
| "loss": 0.1492, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.07559721802237677, | |
| "grad_norm": 0.703405499458313, | |
| "learning_rate": 7.5302419354838715e-06, | |
| "loss": 0.1428, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.017437758290495, | |
| "grad_norm": 0.6589027643203735, | |
| "learning_rate": 9.994400268787099e-06, | |
| "loss": 0.1261, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.0426368309646206, | |
| "grad_norm": 0.6031370759010315, | |
| "learning_rate": 9.71441370814201e-06, | |
| "loss": 0.1202, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.0678359036387461, | |
| "grad_norm": 0.6696850657463074, | |
| "learning_rate": 9.434427147496921e-06, | |
| "loss": 0.1171, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.0833081342606592, | |
| "eval_loss": 0.017358383163809776, | |
| "eval_runtime": 504.91, | |
| "eval_samples_per_second": 0.824, | |
| "eval_steps_per_second": 0.008, | |
| "eval_wer": 0.8985657508208054, | |
| "step": 3307 | |
| }, | |
| { | |
| "epoch": 2.0096764439068644, | |
| "grad_norm": 0.591613233089447, | |
| "learning_rate": 9.154440586851832e-06, | |
| "loss": 0.1042, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.03487551658099, | |
| "grad_norm": 0.5527406930923462, | |
| "learning_rate": 8.874454026206742e-06, | |
| "loss": 0.0861, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.0600745892551156, | |
| "grad_norm": 0.5651601552963257, | |
| "learning_rate": 8.594467465561653e-06, | |
| "loss": 0.0878, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.0019151295232334, | |
| "grad_norm": 0.4751633107662201, | |
| "learning_rate": 8.314480904916565e-06, | |
| "loss": 0.0865, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.027114202197359, | |
| "grad_norm": 0.5415444374084473, | |
| "learning_rate": 8.034494344271475e-06, | |
| "loss": 0.0625, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.0523132748714845, | |
| "grad_norm": 0.5294741988182068, | |
| "learning_rate": 7.754507783626388e-06, | |
| "loss": 0.0636, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.0775123475456105, | |
| "grad_norm": 0.538415789604187, | |
| "learning_rate": 7.474521222981298e-06, | |
| "loss": 0.0652, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.083257736115311, | |
| "eval_loss": 0.01661744900047779, | |
| "eval_runtime": 517.2779, | |
| "eval_samples_per_second": 0.804, | |
| "eval_steps_per_second": 0.008, | |
| "eval_wer": 0.9369659965823781, | |
| "step": 6614 | |
| }, | |
| { | |
| "epoch": 4.019352887813729, | |
| "grad_norm": 0.6536675691604614, | |
| "learning_rate": 7.194534662336209e-06, | |
| "loss": 0.0474, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 4.044551960487854, | |
| "grad_norm": 0.5821442604064941, | |
| "learning_rate": 6.91454810169112e-06, | |
| "loss": 0.0441, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 4.06975103316198, | |
| "grad_norm": 0.5527841448783875, | |
| "learning_rate": 6.6345615410460304e-06, | |
| "loss": 0.0449, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.011591573430098, | |
| "grad_norm": 0.453218549489975, | |
| "learning_rate": 6.354574980400942e-06, | |
| "loss": 0.0369, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 5.036790646104223, | |
| "grad_norm": 0.5013980865478516, | |
| "learning_rate": 6.074588419755852e-06, | |
| "loss": 0.0273, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 5.061989718778349, | |
| "grad_norm": 0.5885359644889832, | |
| "learning_rate": 5.7946018591107636e-06, | |
| "loss": 0.0288, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 5.083207337969963, | |
| "eval_loss": 0.017346344888210297, | |
| "eval_runtime": 530.4397, | |
| "eval_samples_per_second": 0.784, | |
| "eval_steps_per_second": 0.008, | |
| "eval_wer": 1.0060864389532094, | |
| "step": 9921 | |
| }, | |
| { | |
| "epoch": 6.003830259046467, | |
| "grad_norm": 0.4446285665035248, | |
| "learning_rate": 5.514615298465674e-06, | |
| "loss": 0.0276, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 6.029029331720593, | |
| "grad_norm": 0.46283265948295593, | |
| "learning_rate": 5.234628737820585e-06, | |
| "loss": 0.016, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 6.054228404394718, | |
| "grad_norm": 0.42813611030578613, | |
| "learning_rate": 4.954642177175496e-06, | |
| "loss": 0.0173, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 6.079427477068844, | |
| "grad_norm": 0.48960232734680176, | |
| "learning_rate": 4.674655616530407e-06, | |
| "loss": 0.018, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 7.021268017336962, | |
| "grad_norm": 0.3988407254219055, | |
| "learning_rate": 4.394669055885318e-06, | |
| "loss": 0.0109, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 7.046467090011087, | |
| "grad_norm": 0.3698909282684326, | |
| "learning_rate": 4.114682495240229e-06, | |
| "loss": 0.0101, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 7.071666162685213, | |
| "grad_norm": 0.4073663055896759, | |
| "learning_rate": 3.8346959345951395e-06, | |
| "loss": 0.0109, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 7.083156939824614, | |
| "eval_loss": 0.0192726943641901, | |
| "eval_runtime": 518.0983, | |
| "eval_samples_per_second": 0.803, | |
| "eval_steps_per_second": 0.008, | |
| "eval_wer": 0.9907263406485801, | |
| "step": 13228 | |
| }, | |
| { | |
| "epoch": 8.01350670295333, | |
| "grad_norm": 0.33067503571510315, | |
| "learning_rate": 3.5547093739500504e-06, | |
| "loss": 0.0082, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 8.038705775627458, | |
| "grad_norm": 0.3731881380081177, | |
| "learning_rate": 3.2747228133049617e-06, | |
| "loss": 0.0062, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 8.063904848301583, | |
| "grad_norm": 0.2673242688179016, | |
| "learning_rate": 2.9947362526598727e-06, | |
| "loss": 0.0066, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 9.0057453885697, | |
| "grad_norm": 0.18087884783744812, | |
| "learning_rate": 2.7147496920147836e-06, | |
| "loss": 0.0062, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 9.030944461243827, | |
| "grad_norm": 0.250787615776062, | |
| "learning_rate": 2.4347631313696945e-06, | |
| "loss": 0.0038, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 9.056143533917952, | |
| "grad_norm": 0.2255438268184662, | |
| "learning_rate": 2.1547765707246054e-06, | |
| "loss": 0.0041, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 9.081342606592077, | |
| "grad_norm": 0.29366812109947205, | |
| "learning_rate": 1.8747900100795163e-06, | |
| "loss": 0.0044, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 9.083106541679266, | |
| "eval_loss": 0.020848926156759262, | |
| "eval_runtime": 513.0393, | |
| "eval_samples_per_second": 0.811, | |
| "eval_steps_per_second": 0.008, | |
| "eval_wer": 0.9772862546320297, | |
| "step": 16535 | |
| }, | |
| { | |
| "epoch": 10.023183146860196, | |
| "grad_norm": 0.1837795376777649, | |
| "learning_rate": 1.5948034494344272e-06, | |
| "loss": 0.0026, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 10.048382219534322, | |
| "grad_norm": 0.24585728347301483, | |
| "learning_rate": 1.314816888789338e-06, | |
| "loss": 0.0026, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 10.073581292208447, | |
| "grad_norm": 0.15548868477344513, | |
| "learning_rate": 1.0348303281442492e-06, | |
| "loss": 0.0026, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 11.015421832476564, | |
| "grad_norm": 0.08757825195789337, | |
| "learning_rate": 7.554037406204502e-07, | |
| "loss": 0.002, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 11.040620905150691, | |
| "grad_norm": 0.1045205295085907, | |
| "learning_rate": 4.7541717997536123e-07, | |
| "loss": 0.0016, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 11.065819977824816, | |
| "grad_norm": 0.07768367975950241, | |
| "learning_rate": 1.9543061933027217e-07, | |
| "loss": 0.0016, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 11.083056143533918, | |
| "eval_loss": 0.02412882074713707, | |
| "eval_runtime": 518.0169, | |
| "eval_samples_per_second": 0.803, | |
| "eval_steps_per_second": 0.008, | |
| "eval_wer": 0.9926463529366589, | |
| "step": 19842 | |
| }, | |
| { | |
| "epoch": 11.083056143533918, | |
| "step": 19842, | |
| "total_flos": 8.628884758428616e+21, | |
| "train_loss": 0.04815149868992161, | |
| "train_runtime": 124240.955, | |
| "train_samples_per_second": 20.442, | |
| "train_steps_per_second": 0.16 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 19842, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 9223372036854775807, | |
| "save_steps": 3307, | |
| "total_flos": 8.628884758428616e+21, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |