| { | |
| "best_metric": 21.4802, | |
| "best_model_checkpoint": "./tst-translation-output/checkpoint-16000", | |
| "epoch": 2.7845457709711106, | |
| "eval_steps": 2000, | |
| "global_step": 24000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.5441, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 5e-05, | |
| "loss": 1.4175, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.9927274842913664e-05, | |
| "loss": 1.4046, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.985454968582732e-05, | |
| "loss": 1.3784, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "eval_bleu": 18.2602, | |
| "eval_gen_len": 19.2938, | |
| "eval_loss": 1.5514309406280518, | |
| "eval_runtime": 1146.8459, | |
| "eval_samples_per_second": 15.03, | |
| "eval_steps_per_second": 0.94, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.978182452874098e-05, | |
| "loss": 1.3367, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.9709099371654644e-05, | |
| "loss": 1.3112, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 4.9636374214568306e-05, | |
| "loss": 1.2933, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.956364905748196e-05, | |
| "loss": 1.2953, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_bleu": 19.6277, | |
| "eval_gen_len": 18.7905, | |
| "eval_loss": 1.5005673170089722, | |
| "eval_runtime": 1112.1506, | |
| "eval_samples_per_second": 15.499, | |
| "eval_steps_per_second": 0.969, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.9490923900395623e-05, | |
| "loss": 1.2817, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.9418198743309286e-05, | |
| "loss": 1.2722, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.9345473586222954e-05, | |
| "loss": 1.2491, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 4.927274842913661e-05, | |
| "loss": 1.2446, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_bleu": 20.2667, | |
| "eval_gen_len": 19.2503, | |
| "eval_loss": 1.4663728475570679, | |
| "eval_runtime": 1107.4417, | |
| "eval_samples_per_second": 15.565, | |
| "eval_steps_per_second": 0.973, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 4.920002327205027e-05, | |
| "loss": 1.2323, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 4.9127298114963934e-05, | |
| "loss": 1.2242, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 4.905457295787759e-05, | |
| "loss": 1.2223, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 4.898184780079125e-05, | |
| "loss": 1.2095, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "eval_bleu": 20.8962, | |
| "eval_gen_len": 18.9352, | |
| "eval_loss": 1.448201298713684, | |
| "eval_runtime": 1086.8921, | |
| "eval_samples_per_second": 15.859, | |
| "eval_steps_per_second": 0.992, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.8909122643704913e-05, | |
| "loss": 1.2022, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.8836397486618576e-05, | |
| "loss": 0.9756, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 4.876367232953223e-05, | |
| "loss": 0.9306, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.869094717244589e-05, | |
| "loss": 0.9279, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "eval_bleu": 20.9876, | |
| "eval_gen_len": 19.093, | |
| "eval_loss": 1.4799305200576782, | |
| "eval_runtime": 1094.4619, | |
| "eval_samples_per_second": 15.749, | |
| "eval_steps_per_second": 0.985, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 4.8618222015359555e-05, | |
| "loss": 0.9366, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 4.854549685827322e-05, | |
| "loss": 0.9408, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 4.847277170118687e-05, | |
| "loss": 0.9456, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.8400046544100535e-05, | |
| "loss": 0.9604, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_bleu": 21.261, | |
| "eval_gen_len": 18.8735, | |
| "eval_loss": 1.4671677350997925, | |
| "eval_runtime": 1083.8306, | |
| "eval_samples_per_second": 15.904, | |
| "eval_steps_per_second": 0.995, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 4.83273213870142e-05, | |
| "loss": 0.9375, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 4.825459622992786e-05, | |
| "loss": 0.9412, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 4.818187107284152e-05, | |
| "loss": 0.9441, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 4.810914591575518e-05, | |
| "loss": 0.9543, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "eval_bleu": 21.1987, | |
| "eval_gen_len": 18.8396, | |
| "eval_loss": 1.4610902070999146, | |
| "eval_runtime": 1087.5897, | |
| "eval_samples_per_second": 15.849, | |
| "eval_steps_per_second": 0.991, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "learning_rate": 4.8036420758668845e-05, | |
| "loss": 0.9561, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 4.79636956015825e-05, | |
| "loss": 0.9587, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 4.789097044449616e-05, | |
| "loss": 0.9553, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 4.7818245287409825e-05, | |
| "loss": 0.9532, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "eval_bleu": 21.4802, | |
| "eval_gen_len": 18.8239, | |
| "eval_loss": 1.4428682327270508, | |
| "eval_runtime": 1090.8288, | |
| "eval_samples_per_second": 15.802, | |
| "eval_steps_per_second": 0.988, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "learning_rate": 4.7745520130323487e-05, | |
| "loss": 0.9725, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 4.767279497323714e-05, | |
| "loss": 0.9598, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 4.7600069816150804e-05, | |
| "loss": 0.8071, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 4.7527344659064466e-05, | |
| "loss": 0.6681, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "eval_bleu": 21.1981, | |
| "eval_gen_len": 18.6116, | |
| "eval_loss": 1.544991135597229, | |
| "eval_runtime": 1065.2077, | |
| "eval_samples_per_second": 16.182, | |
| "eval_steps_per_second": 1.012, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "learning_rate": 4.745461950197813e-05, | |
| "loss": 0.6705, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 4.7381894344891784e-05, | |
| "loss": 0.6931, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "learning_rate": 4.7309169187805446e-05, | |
| "loss": 0.697, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 4.723644403071911e-05, | |
| "loss": 0.6971, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "eval_bleu": 21.3101, | |
| "eval_gen_len": 18.892, | |
| "eval_loss": 1.5515910387039185, | |
| "eval_runtime": 1074.5394, | |
| "eval_samples_per_second": 16.041, | |
| "eval_steps_per_second": 1.003, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 4.716371887363277e-05, | |
| "loss": 0.7071, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 4.7090993716546425e-05, | |
| "loss": 0.7177, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "learning_rate": 4.701826855946009e-05, | |
| "loss": 0.7177, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.6945543402373756e-05, | |
| "loss": 0.7283, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "eval_bleu": 20.902, | |
| "eval_gen_len": 18.6448, | |
| "eval_loss": 1.5404880046844482, | |
| "eval_runtime": 1076.1481, | |
| "eval_samples_per_second": 16.017, | |
| "eval_steps_per_second": 1.002, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "learning_rate": 4.687281824528741e-05, | |
| "loss": 0.7257, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 4.6800093088201073e-05, | |
| "loss": 0.7241, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 4.6727367931114736e-05, | |
| "loss": 0.7339, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.66546427740284e-05, | |
| "loss": 0.7308, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "eval_bleu": 21.3017, | |
| "eval_gen_len": 18.2578, | |
| "eval_loss": 1.5363101959228516, | |
| "eval_runtime": 1054.8574, | |
| "eval_samples_per_second": 16.341, | |
| "eval_steps_per_second": 1.022, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "step": 24000, | |
| "total_flos": 8.32177830887424e+17, | |
| "train_loss": 1.005988956451416, | |
| "train_runtime": 36356.2838, | |
| "train_samples_per_second": 151.721, | |
| "train_steps_per_second": 9.483 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 344760, | |
| "num_train_epochs": 40, | |
| "save_steps": 2000, | |
| "total_flos": 8.32177830887424e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |