| { | |
| "best_metric": 27.8224, | |
| "best_model_checkpoint": "./ko-en_mbartLarge_exp20p_batch/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": 1.5446, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.343, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 1.2767, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 5e-05, | |
| "loss": 1.242, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.9995663025798074e-05, | |
| "loss": 1.1213, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.998261883224581e-05, | |
| "loss": 1.0725, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 4.996087194586878e-05, | |
| "loss": 1.0628, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 4.99304299421927e-05, | |
| "loss": 1.0603, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "eval_bleu": 27.7215, | |
| "eval_gen_len": 18.8456, | |
| "eval_loss": 1.1473960876464844, | |
| "eval_runtime": 1471.0712, | |
| "eval_samples_per_second": 11.717, | |
| "eval_steps_per_second": 1.465, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 4.989130342568659e-05, | |
| "loss": 0.9588, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 4.984371446700271e-05, | |
| "loss": 0.8217, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.978741887280011e-05, | |
| "loss": 0.8436, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.972238441511973e-05, | |
| "loss": 0.8488, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 4.964873791724829e-05, | |
| "loss": 0.8471, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 4.956650503393613e-05, | |
| "loss": 0.6385, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.947571441100011e-05, | |
| "loss": 0.6568, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 4.937660479537447e-05, | |
| "loss": 0.6774, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "eval_bleu": 27.8224, | |
| "eval_gen_len": 18.6267, | |
| "eval_loss": 1.2035210132598877, | |
| "eval_runtime": 1442.7545, | |
| "eval_samples_per_second": 11.947, | |
| "eval_steps_per_second": 1.494, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.94, | |
| "learning_rate": 4.9269037523258574e-05, | |
| "loss": 0.6888, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "learning_rate": 4.915304991597116e-05, | |
| "loss": 0.5341, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.41, | |
| "learning_rate": 4.902842459717919e-05, | |
| "loss": 0.505, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 4.889570329351636e-05, | |
| "loss": 0.5258, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "learning_rate": 4.8754400304796405e-05, | |
| "loss": 0.5402, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "learning_rate": 4.860482247305548e-05, | |
| "loss": 0.4654, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 5.34, | |
| "learning_rate": 4.8447021903716625e-05, | |
| "loss": 0.3817, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "learning_rate": 4.8281053566592745e-05, | |
| "loss": 0.4006, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "eval_bleu": 27.049, | |
| "eval_gen_len": 18.5618, | |
| "eval_loss": 1.4320647716522217, | |
| "eval_runtime": 1443.7655, | |
| "eval_samples_per_second": 11.939, | |
| "eval_steps_per_second": 1.493, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "learning_rate": 4.810697527673781e-05, | |
| "loss": 0.4168, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "learning_rate": 4.792484767430706e-05, | |
| "loss": 0.4084, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 6.27, | |
| "learning_rate": 4.7734734203432976e-05, | |
| "loss": 0.2881, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "learning_rate": 4.7537105014759254e-05, | |
| "loss": 0.3053, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "learning_rate": 4.733123687447865e-05, | |
| "loss": 0.3213, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 6.96, | |
| "learning_rate": 4.711758965001871e-05, | |
| "loss": 0.3334, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 7.19, | |
| "learning_rate": 4.689623776536918e-05, | |
| "loss": 0.2383, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 7.43, | |
| "learning_rate": 4.66672583284374e-05, | |
| "loss": 0.2328, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.43, | |
| "eval_bleu": 25.8623, | |
| "eval_gen_len": 18.4608, | |
| "eval_loss": 1.73158597946167, | |
| "eval_runtime": 1439.3983, | |
| "eval_samples_per_second": 11.975, | |
| "eval_steps_per_second": 1.497, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.66, | |
| "learning_rate": 4.643073110418783e-05, | |
| "loss": 0.248, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 7.89, | |
| "learning_rate": 4.618673848685586e-05, | |
| "loss": 0.257, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 8.12, | |
| "learning_rate": 4.593536547124587e-05, | |
| "loss": 0.2129, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 8.35, | |
| "learning_rate": 4.567722417359549e-05, | |
| "loss": 0.179, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 8.59, | |
| "learning_rate": 4.541136991303665e-05, | |
| "loss": 0.1914, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 8.82, | |
| "learning_rate": 4.513840535376658e-05, | |
| "loss": 0.2002, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "learning_rate": 4.485842558296277e-05, | |
| "loss": 0.1893, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 9.28, | |
| "learning_rate": 4.457152813155099e-05, | |
| "loss": 0.1369, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.28, | |
| "eval_bleu": 25.6902, | |
| "eval_gen_len": 18.3337, | |
| "eval_loss": 1.9367432594299316, | |
| "eval_runtime": 1429.8923, | |
| "eval_samples_per_second": 12.055, | |
| "eval_steps_per_second": 1.507, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.51, | |
| "learning_rate": 4.4277812940230354e-05, | |
| "loss": 0.1495, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 9.75, | |
| "learning_rate": 4.3977989818708614e-05, | |
| "loss": 0.1579, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 9.98, | |
| "learning_rate": 4.367096154941818e-05, | |
| "loss": 0.1631, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 10.21, | |
| "learning_rate": 4.3357429252499605e-05, | |
| "loss": 0.1143, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 10.44, | |
| "learning_rate": 4.3037502146897856e-05, | |
| "loss": 0.1177, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 10.67, | |
| "learning_rate": 4.2711291679188836e-05, | |
| "loss": 0.1267, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 10.91, | |
| "learning_rate": 4.2378911484757056e-05, | |
| "loss": 0.1315, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "learning_rate": 4.204047734821069e-05, | |
| "loss": 0.1067, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "eval_bleu": 26.0105, | |
| "eval_gen_len": 18.4579, | |
| "eval_loss": 2.086793899536133, | |
| "eval_runtime": 1436.0818, | |
| "eval_samples_per_second": 12.003, | |
| "eval_steps_per_second": 1.501, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "step": 24000, | |
| "total_flos": 3.328711323549696e+18, | |
| "train_loss": 0.5267434423764547, | |
| "train_runtime": 55597.7197, | |
| "train_samples_per_second": 99.213, | |
| "train_steps_per_second": 1.55 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 86160, | |
| "num_train_epochs": 40, | |
| "save_steps": 4000, | |
| "total_flos": 3.328711323549696e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |