| { | |
| "best_metric": 0.865190327167511, | |
| "best_model_checkpoint": "./mbart-mmt_mid3_ko-ja/checkpoint-25500", | |
| "epoch": 4.782146652497343, | |
| "eval_steps": 1500, | |
| "global_step": 31500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 5e-05, | |
| "loss": 3.3601, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.98913256102067e-05, | |
| "loss": 2.106, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.97826512204134e-05, | |
| "loss": 1.6216, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "eval_bleu": 2.686, | |
| "eval_gen_len": 17.599, | |
| "eval_loss": 1.522924780845642, | |
| "eval_runtime": 866.1912, | |
| "eval_samples_per_second": 15.207, | |
| "eval_steps_per_second": 0.951, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.96739768306201e-05, | |
| "loss": 1.508, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 4.95653024408268e-05, | |
| "loss": 1.4292, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.945662805103349e-05, | |
| "loss": 1.3587, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_bleu": 4.0749, | |
| "eval_gen_len": 17.3772, | |
| "eval_loss": 1.306133508682251, | |
| "eval_runtime": 848.4414, | |
| "eval_samples_per_second": 15.525, | |
| "eval_steps_per_second": 0.971, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 4.9347953661240194e-05, | |
| "loss": 1.3063, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 4.923927927144689e-05, | |
| "loss": 1.2722, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 4.9130604881653594e-05, | |
| "loss": 1.2279, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "eval_bleu": 5.2878, | |
| "eval_gen_len": 17.3642, | |
| "eval_loss": 1.1880881786346436, | |
| "eval_runtime": 848.9163, | |
| "eval_samples_per_second": 15.516, | |
| "eval_steps_per_second": 0.971, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 4.902193049186029e-05, | |
| "loss": 1.203, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.891325610206699e-05, | |
| "loss": 1.1744, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 4.880458171227369e-05, | |
| "loss": 1.1408, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "eval_bleu": 5.4783, | |
| "eval_gen_len": 17.4093, | |
| "eval_loss": 1.0994044542312622, | |
| "eval_runtime": 903.2669, | |
| "eval_samples_per_second": 14.583, | |
| "eval_steps_per_second": 0.912, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.8695907322480386e-05, | |
| "loss": 1.1196, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 4.858723293268709e-05, | |
| "loss": 1.014, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 4.847855854289378e-05, | |
| "loss": 0.9977, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "eval_bleu": 7.6015, | |
| "eval_gen_len": 17.36, | |
| "eval_loss": 1.0312957763671875, | |
| "eval_runtime": 872.8057, | |
| "eval_samples_per_second": 15.092, | |
| "eval_steps_per_second": 0.944, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 4.836988415310048e-05, | |
| "loss": 0.9903, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 4.826120976330718e-05, | |
| "loss": 0.9752, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 4.815253537351388e-05, | |
| "loss": 0.9582, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "eval_bleu": 8.2303, | |
| "eval_gen_len": 17.3526, | |
| "eval_loss": 0.9917855262756348, | |
| "eval_runtime": 851.0578, | |
| "eval_samples_per_second": 15.477, | |
| "eval_steps_per_second": 0.968, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 4.804386098372058e-05, | |
| "loss": 0.9684, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "learning_rate": 4.793518659392728e-05, | |
| "loss": 0.9641, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "learning_rate": 4.782651220413398e-05, | |
| "loss": 0.9525, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "eval_bleu": 8.2837, | |
| "eval_gen_len": 17.2597, | |
| "eval_loss": 0.981093168258667, | |
| "eval_runtime": 868.378, | |
| "eval_samples_per_second": 15.169, | |
| "eval_steps_per_second": 0.949, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 4.7717837814340674e-05, | |
| "loss": 0.943, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 4.760916342454738e-05, | |
| "loss": 0.936, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 4.750048903475407e-05, | |
| "loss": 0.9415, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "eval_bleu": 8.1592, | |
| "eval_gen_len": 17.2241, | |
| "eval_loss": 0.9589386582374573, | |
| "eval_runtime": 853.0091, | |
| "eval_samples_per_second": 15.442, | |
| "eval_steps_per_second": 0.966, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 4.739181464496077e-05, | |
| "loss": 0.9327, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 4.7283140255167467e-05, | |
| "loss": 0.9219, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 4.717446586537417e-05, | |
| "loss": 0.856, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "eval_bleu": 7.8401, | |
| "eval_gen_len": 17.4066, | |
| "eval_loss": 0.9462171196937561, | |
| "eval_runtime": 870.9328, | |
| "eval_samples_per_second": 15.124, | |
| "eval_steps_per_second": 0.946, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 4.7065791475580866e-05, | |
| "loss": 0.816, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 4.695711708578757e-05, | |
| "loss": 0.819, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 4.6848442695994266e-05, | |
| "loss": 0.8273, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "eval_bleu": 8.6082, | |
| "eval_gen_len": 17.1918, | |
| "eval_loss": 0.9335508942604065, | |
| "eval_runtime": 847.8053, | |
| "eval_samples_per_second": 15.537, | |
| "eval_steps_per_second": 0.972, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "learning_rate": 4.673976830620096e-05, | |
| "loss": 0.8154, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 4.6631093916407665e-05, | |
| "loss": 0.8281, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 4.6522419526614355e-05, | |
| "loss": 0.8066, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_bleu": 9.7751, | |
| "eval_gen_len": 17.5198, | |
| "eval_loss": 0.921973466873169, | |
| "eval_runtime": 888.35, | |
| "eval_samples_per_second": 14.827, | |
| "eval_steps_per_second": 0.928, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "learning_rate": 4.641374513682106e-05, | |
| "loss": 0.812, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 4.6305070747027754e-05, | |
| "loss": 0.8117, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 4.619639635723446e-05, | |
| "loss": 0.784, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "eval_bleu": 10.292, | |
| "eval_gen_len": 17.4097, | |
| "eval_loss": 0.8948560357093811, | |
| "eval_runtime": 887.752, | |
| "eval_samples_per_second": 14.837, | |
| "eval_steps_per_second": 0.928, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 4.6087721967441154e-05, | |
| "loss": 0.8113, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "learning_rate": 4.597904757764786e-05, | |
| "loss": 0.7942, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 4.5870373187854554e-05, | |
| "loss": 0.8016, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "eval_bleu": 9.0262, | |
| "eval_gen_len": 17.4097, | |
| "eval_loss": 0.8958194255828857, | |
| "eval_runtime": 868.3186, | |
| "eval_samples_per_second": 15.17, | |
| "eval_steps_per_second": 0.949, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 4.576169879806125e-05, | |
| "loss": 0.7417, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.11, | |
| "learning_rate": 4.5653024408267946e-05, | |
| "loss": 0.683, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 4.554435001847465e-05, | |
| "loss": 0.6872, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "eval_bleu": 9.7549, | |
| "eval_gen_len": 17.2672, | |
| "eval_loss": 0.9043093919754028, | |
| "eval_runtime": 855.9409, | |
| "eval_samples_per_second": 15.389, | |
| "eval_steps_per_second": 0.963, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "learning_rate": 4.5435675628681346e-05, | |
| "loss": 0.6928, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 3.34, | |
| "learning_rate": 4.532700123888804e-05, | |
| "loss": 0.6973, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "learning_rate": 4.5218326849094746e-05, | |
| "loss": 0.7107, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "eval_bleu": 10.3016, | |
| "eval_gen_len": 17.0973, | |
| "eval_loss": 0.8993721008300781, | |
| "eval_runtime": 837.6757, | |
| "eval_samples_per_second": 15.724, | |
| "eval_steps_per_second": 0.984, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 3.49, | |
| "learning_rate": 4.510965245930144e-05, | |
| "loss": 0.6899, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 4.5000978069508145e-05, | |
| "loss": 0.6794, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 4.489230367971484e-05, | |
| "loss": 0.6726, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "eval_bleu": 10.5183, | |
| "eval_gen_len": 17.2871, | |
| "eval_loss": 0.8746942281723022, | |
| "eval_runtime": 850.8583, | |
| "eval_samples_per_second": 15.481, | |
| "eval_steps_per_second": 0.968, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.72, | |
| "learning_rate": 4.4783629289921545e-05, | |
| "loss": 0.6782, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "learning_rate": 4.4674954900128234e-05, | |
| "loss": 0.6821, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "learning_rate": 4.456628051033494e-05, | |
| "loss": 0.6699, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "eval_bleu": 10.1883, | |
| "eval_gen_len": 17.2057, | |
| "eval_loss": 0.865190327167511, | |
| "eval_runtime": 852.5269, | |
| "eval_samples_per_second": 15.451, | |
| "eval_steps_per_second": 0.967, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "learning_rate": 4.4457606120541634e-05, | |
| "loss": 0.7029, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "learning_rate": 4.434893173074833e-05, | |
| "loss": 0.6761, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "learning_rate": 4.4240257340955033e-05, | |
| "loss": 0.612, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "eval_bleu": 9.5697, | |
| "eval_gen_len": 17.2443, | |
| "eval_loss": 0.8949306607246399, | |
| "eval_runtime": 850.694, | |
| "eval_samples_per_second": 15.484, | |
| "eval_steps_per_second": 0.969, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "learning_rate": 4.413158295116173e-05, | |
| "loss": 0.6119, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "learning_rate": 4.402290856136843e-05, | |
| "loss": 0.6094, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 4.391423417157513e-05, | |
| "loss": 0.621, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "eval_bleu": 10.8592, | |
| "eval_gen_len": 17.329, | |
| "eval_loss": 0.8904227614402771, | |
| "eval_runtime": 868.3516, | |
| "eval_samples_per_second": 15.169, | |
| "eval_steps_per_second": 0.949, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "learning_rate": 4.380555978178183e-05, | |
| "loss": 0.6302, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 4.48, | |
| "learning_rate": 4.369688539198852e-05, | |
| "loss": 0.6206, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 4.55, | |
| "learning_rate": 4.3588211002195225e-05, | |
| "loss": 0.6219, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.55, | |
| "eval_bleu": 10.925, | |
| "eval_gen_len": 17.482, | |
| "eval_loss": 0.8771834969520569, | |
| "eval_runtime": 864.4075, | |
| "eval_samples_per_second": 15.238, | |
| "eval_steps_per_second": 0.953, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.63, | |
| "learning_rate": 4.347953661240192e-05, | |
| "loss": 0.6166, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 4.71, | |
| "learning_rate": 4.337086222260862e-05, | |
| "loss": 0.6057, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "learning_rate": 4.326218783281532e-05, | |
| "loss": 0.6164, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "eval_bleu": 11.8749, | |
| "eval_gen_len": 17.1624, | |
| "eval_loss": 0.8693726658821106, | |
| "eval_runtime": 846.6139, | |
| "eval_samples_per_second": 15.558, | |
| "eval_steps_per_second": 0.973, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "step": 31500, | |
| "total_flos": 1.0923720993448919e+18, | |
| "train_loss": 0.9323235357375372, | |
| "train_runtime": 42332.75, | |
| "train_samples_per_second": 87.131, | |
| "train_steps_per_second": 5.446 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 230545, | |
| "num_train_epochs": 35, | |
| "save_steps": 1500, | |
| "total_flos": 1.0923720993448919e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |