| { | |
| "best_metric": 27.8203, | |
| "best_model_checkpoint": "./ko-en_mbartLarge_exp20p_batch64_linear/checkpoint-8000", | |
| "epoch": 11.138183083884442, | |
| "eval_steps": 4000, | |
| "global_step": 24000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 1.25e-05, | |
| "loss": 1.5426, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.343, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 1.2768, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 5e-05, | |
| "loss": 1.242, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.970294676806084e-05, | |
| "loss": 1.1208, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.940589353612168e-05, | |
| "loss": 1.0718, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 4.910884030418251e-05, | |
| "loss": 1.061, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 4.881178707224335e-05, | |
| "loss": 1.0585, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "eval_bleu": 27.494, | |
| "eval_gen_len": 18.9136, | |
| "eval_loss": 1.1434648036956787, | |
| "eval_runtime": 1314.8151, | |
| "eval_samples_per_second": 13.11, | |
| "eval_steps_per_second": 1.639, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 4.851473384030418e-05, | |
| "loss": 0.9558, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 4.8217680608365025e-05, | |
| "loss": 0.8186, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.792062737642586e-05, | |
| "loss": 0.8398, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.7623574144486695e-05, | |
| "loss": 0.8454, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 4.732652091254753e-05, | |
| "loss": 0.8416, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 4.702946768060837e-05, | |
| "loss": 0.6354, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.67324144486692e-05, | |
| "loss": 0.6545, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 4.643536121673004e-05, | |
| "loss": 0.6719, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "eval_bleu": 27.8203, | |
| "eval_gen_len": 18.6815, | |
| "eval_loss": 1.204200267791748, | |
| "eval_runtime": 1295.2995, | |
| "eval_samples_per_second": 13.307, | |
| "eval_steps_per_second": 1.664, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.94, | |
| "learning_rate": 4.613830798479088e-05, | |
| "loss": 0.6824, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "learning_rate": 4.584125475285171e-05, | |
| "loss": 0.5317, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.41, | |
| "learning_rate": 4.554420152091255e-05, | |
| "loss": 0.5004, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 4.524714828897338e-05, | |
| "loss": 0.5208, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "learning_rate": 4.4950095057034226e-05, | |
| "loss": 0.5339, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "learning_rate": 4.465304182509506e-05, | |
| "loss": 0.4608, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 5.34, | |
| "learning_rate": 4.4355988593155896e-05, | |
| "loss": 0.3783, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "learning_rate": 4.405893536121673e-05, | |
| "loss": 0.3964, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "eval_bleu": 27.1711, | |
| "eval_gen_len": 18.5184, | |
| "eval_loss": 1.4261032342910767, | |
| "eval_runtime": 1285.7174, | |
| "eval_samples_per_second": 13.407, | |
| "eval_steps_per_second": 1.676, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "learning_rate": 4.376188212927757e-05, | |
| "loss": 0.4104, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "learning_rate": 4.3464828897338404e-05, | |
| "loss": 0.4036, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 6.27, | |
| "learning_rate": 4.316777566539924e-05, | |
| "loss": 0.2859, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "learning_rate": 4.2870722433460074e-05, | |
| "loss": 0.3004, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "learning_rate": 4.257366920152092e-05, | |
| "loss": 0.3148, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 6.96, | |
| "learning_rate": 4.227661596958175e-05, | |
| "loss": 0.3262, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 7.19, | |
| "learning_rate": 4.197956273764259e-05, | |
| "loss": 0.234, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 7.43, | |
| "learning_rate": 4.168250950570343e-05, | |
| "loss": 0.2282, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.43, | |
| "eval_bleu": 26.1891, | |
| "eval_gen_len": 18.5479, | |
| "eval_loss": 1.7274521589279175, | |
| "eval_runtime": 1283.8263, | |
| "eval_samples_per_second": 13.426, | |
| "eval_steps_per_second": 1.679, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.66, | |
| "learning_rate": 4.138545627376426e-05, | |
| "loss": 0.2426, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 7.89, | |
| "learning_rate": 4.10884030418251e-05, | |
| "loss": 0.2507, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 8.12, | |
| "learning_rate": 4.079134980988593e-05, | |
| "loss": 0.2085, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 8.35, | |
| "learning_rate": 4.0494296577946774e-05, | |
| "loss": 0.1756, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 8.59, | |
| "learning_rate": 4.0197243346007605e-05, | |
| "loss": 0.1859, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 8.82, | |
| "learning_rate": 3.9900190114068444e-05, | |
| "loss": 0.1957, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "learning_rate": 3.9603136882129275e-05, | |
| "loss": 0.1845, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 9.28, | |
| "learning_rate": 3.930608365019012e-05, | |
| "loss": 0.1338, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.28, | |
| "eval_bleu": 26.1533, | |
| "eval_gen_len": 18.505, | |
| "eval_loss": 1.9251435995101929, | |
| "eval_runtime": 1285.914, | |
| "eval_samples_per_second": 13.404, | |
| "eval_steps_per_second": 1.676, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.51, | |
| "learning_rate": 3.900903041825095e-05, | |
| "loss": 0.1447, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 9.75, | |
| "learning_rate": 3.871197718631179e-05, | |
| "loss": 0.1527, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 9.98, | |
| "learning_rate": 3.841492395437263e-05, | |
| "loss": 0.1584, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 10.21, | |
| "learning_rate": 3.811787072243346e-05, | |
| "loss": 0.1119, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 10.44, | |
| "learning_rate": 3.78208174904943e-05, | |
| "loss": 0.1127, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 10.67, | |
| "learning_rate": 3.752376425855513e-05, | |
| "loss": 0.1233, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 10.91, | |
| "learning_rate": 3.7226711026615975e-05, | |
| "loss": 0.1269, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "learning_rate": 3.6929657794676806e-05, | |
| "loss": 0.1033, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "eval_bleu": 26.0643, | |
| "eval_gen_len": 18.4275, | |
| "eval_loss": 2.0758345127105713, | |
| "eval_runtime": 1283.6981, | |
| "eval_samples_per_second": 13.428, | |
| "eval_steps_per_second": 1.679, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "step": 24000, | |
| "total_flos": 3.328711323549696e+18, | |
| "train_loss": 0.5229005074501037, | |
| "train_runtime": 88579.3039, | |
| "train_samples_per_second": 62.272, | |
| "train_steps_per_second": 0.973 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 86160, | |
| "num_train_epochs": 40, | |
| "save_steps": 4000, | |
| "total_flos": 3.328711323549696e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |