| { | |
| "best_global_step": 10000, | |
| "best_metric": 63.44155467990254, | |
| "best_model_checkpoint": "nllb-200-600M-dzo-eng-30k-checkpoints/checkpoint-10000", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 12500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.2576698958873749, | |
| "learning_rate": 2.88024e-05, | |
| "loss": 0.0753, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 0.19992510974407196, | |
| "learning_rate": 2.7602400000000002e-05, | |
| "loss": 0.0735, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 0.2977311909198761, | |
| "learning_rate": 2.64024e-05, | |
| "loss": 0.0713, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.21389248967170715, | |
| "learning_rate": 2.52024e-05, | |
| "loss": 0.071, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.2741245627403259, | |
| "learning_rate": 2.4002400000000002e-05, | |
| "loss": 0.0688, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_bleu": 62.03431583845565, | |
| "eval_loss": 0.06440743803977966, | |
| "eval_runtime": 243.3832, | |
| "eval_samples_per_second": 6.163, | |
| "eval_steps_per_second": 1.541, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.24694935977458954, | |
| "learning_rate": 2.28024e-05, | |
| "loss": 0.0553, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 0.26646921038627625, | |
| "learning_rate": 2.1602400000000002e-05, | |
| "loss": 0.0563, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 0.25563061237335205, | |
| "learning_rate": 2.04024e-05, | |
| "loss": 0.0538, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 0.25169041752815247, | |
| "learning_rate": 1.92024e-05, | |
| "loss": 0.0546, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.2201855629682541, | |
| "learning_rate": 1.80024e-05, | |
| "loss": 0.0543, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_bleu": 62.84538472555695, | |
| "eval_loss": 0.06339649111032486, | |
| "eval_runtime": 245.6581, | |
| "eval_samples_per_second": 6.106, | |
| "eval_steps_per_second": 1.527, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 0.20503699779510498, | |
| "learning_rate": 1.68024e-05, | |
| "loss": 0.0436, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 0.18902115523815155, | |
| "learning_rate": 1.56024e-05, | |
| "loss": 0.0444, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "grad_norm": 0.39264440536499023, | |
| "learning_rate": 1.4402400000000001e-05, | |
| "loss": 0.0434, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 0.22416116297245026, | |
| "learning_rate": 1.3202400000000001e-05, | |
| "loss": 0.0443, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.2807520925998688, | |
| "learning_rate": 1.20024e-05, | |
| "loss": 0.0444, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_bleu": 63.39396750181061, | |
| "eval_loss": 0.06433024257421494, | |
| "eval_runtime": 245.8516, | |
| "eval_samples_per_second": 6.101, | |
| "eval_steps_per_second": 1.525, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "grad_norm": 0.2532879710197449, | |
| "learning_rate": 1.08024e-05, | |
| "loss": 0.0371, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "grad_norm": 0.20367200672626495, | |
| "learning_rate": 9.6024e-06, | |
| "loss": 0.036, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "grad_norm": 0.24893485009670258, | |
| "learning_rate": 8.4024e-06, | |
| "loss": 0.0364, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "grad_norm": 0.27603277564048767, | |
| "learning_rate": 7.2024e-06, | |
| "loss": 0.0366, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.21856920421123505, | |
| "learning_rate": 6.0024e-06, | |
| "loss": 0.0377, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_bleu": 63.44155467990254, | |
| "eval_loss": 0.06584766507148743, | |
| "eval_runtime": 246.2205, | |
| "eval_samples_per_second": 6.092, | |
| "eval_steps_per_second": 1.523, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.2, | |
| "grad_norm": 0.21814288198947906, | |
| "learning_rate": 4.8024e-06, | |
| "loss": 0.0328, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "grad_norm": 0.20107074081897736, | |
| "learning_rate": 3.6024000000000004e-06, | |
| "loss": 0.0321, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 4.6, | |
| "grad_norm": 0.2721525728702545, | |
| "learning_rate": 2.4024e-06, | |
| "loss": 0.0323, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "grad_norm": 0.33091649413108826, | |
| "learning_rate": 1.2024e-06, | |
| "loss": 0.0329, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.2466353476047516, | |
| "learning_rate": 2.4000000000000004e-09, | |
| "loss": 0.0315, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_bleu": 63.062347513148154, | |
| "eval_loss": 0.06781521439552307, | |
| "eval_runtime": 251.1652, | |
| "eval_samples_per_second": 5.972, | |
| "eval_steps_per_second": 1.493, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 12500, | |
| "total_flos": 2.70888075264e+16, | |
| "train_loss": 0.04798975685119629, | |
| "train_runtime": 11153.1781, | |
| "train_samples_per_second": 8.966, | |
| "train_steps_per_second": 1.121 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 12500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.70888075264e+16, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |