| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "eval_steps": 500, | |
| "global_step": 1320, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_bleu": 0.2664373993158469, | |
| "eval_loss": 7.194581985473633, | |
| "eval_runtime": 8.3226, | |
| "eval_samples_per_second": 5.046, | |
| "eval_steps_per_second": 0.36, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_bleu": 0.1319619723907014, | |
| "eval_loss": 5.048046588897705, | |
| "eval_runtime": 6.4117, | |
| "eval_samples_per_second": 6.551, | |
| "eval_steps_per_second": 0.468, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_bleu": 0.06949242308398045, | |
| "eval_loss": 2.2946908473968506, | |
| "eval_runtime": 13.8186, | |
| "eval_samples_per_second": 3.039, | |
| "eval_steps_per_second": 0.217, | |
| "step": 264 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_bleu": 0.0, | |
| "eval_loss": 1.7485623359680176, | |
| "eval_runtime": 13.9926, | |
| "eval_samples_per_second": 3.002, | |
| "eval_steps_per_second": 0.214, | |
| "step": 352 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_bleu": 0.0, | |
| "eval_loss": 1.7318042516708374, | |
| "eval_runtime": 14.0045, | |
| "eval_samples_per_second": 2.999, | |
| "eval_steps_per_second": 0.214, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 5.681818181818182, | |
| "grad_norm": 0.5759038925170898, | |
| "learning_rate": 4.96e-05, | |
| "loss": 4.2283, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_bleu": 0.011387252994475159, | |
| "eval_loss": 1.734961986541748, | |
| "eval_runtime": 13.9988, | |
| "eval_samples_per_second": 3.0, | |
| "eval_steps_per_second": 0.214, | |
| "step": 528 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_bleu": 0.061716390858398085, | |
| "eval_loss": 1.7283143997192383, | |
| "eval_runtime": 13.8819, | |
| "eval_samples_per_second": 3.026, | |
| "eval_steps_per_second": 0.216, | |
| "step": 616 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_bleu": 4.8538185401433145, | |
| "eval_loss": 1.7294987440109253, | |
| "eval_runtime": 9.3071, | |
| "eval_samples_per_second": 4.513, | |
| "eval_steps_per_second": 0.322, | |
| "step": 704 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_bleu": 1.442462891266846, | |
| "eval_loss": 1.7348395586013794, | |
| "eval_runtime": 12.56, | |
| "eval_samples_per_second": 3.344, | |
| "eval_steps_per_second": 0.239, | |
| "step": 792 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_bleu": 5.259829291051203, | |
| "eval_loss": 1.7377452850341797, | |
| "eval_runtime": 7.4674, | |
| "eval_samples_per_second": 5.624, | |
| "eval_steps_per_second": 0.402, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_bleu": 5.116295525045743, | |
| "eval_loss": 1.7416119575500488, | |
| "eval_runtime": 9.8057, | |
| "eval_samples_per_second": 4.283, | |
| "eval_steps_per_second": 0.306, | |
| "step": 968 | |
| }, | |
| { | |
| "epoch": 11.363636363636363, | |
| "grad_norm": 0.17466464638710022, | |
| "learning_rate": 1.975609756097561e-05, | |
| "loss": 1.564, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_bleu": 5.586956856799072, | |
| "eval_loss": 1.7407861948013306, | |
| "eval_runtime": 7.391, | |
| "eval_samples_per_second": 5.683, | |
| "eval_steps_per_second": 0.406, | |
| "step": 1056 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_bleu": 5.6086622973490075, | |
| "eval_loss": 1.7451168298721313, | |
| "eval_runtime": 4.8639, | |
| "eval_samples_per_second": 8.635, | |
| "eval_steps_per_second": 0.617, | |
| "step": 1144 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_bleu": 5.585094069359239, | |
| "eval_loss": 1.7451788187026978, | |
| "eval_runtime": 7.4561, | |
| "eval_samples_per_second": 5.633, | |
| "eval_steps_per_second": 0.402, | |
| "step": 1232 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_bleu": 5.572328628563582, | |
| "eval_loss": 1.7453956604003906, | |
| "eval_runtime": 7.2527, | |
| "eval_samples_per_second": 5.791, | |
| "eval_steps_per_second": 0.414, | |
| "step": 1320 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1320, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "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": 1.140167908786176e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |