{ "best_global_step": 241, "best_metric": 1.9624056816101074, "best_model_checkpoint": "./mt5_base/indic_loss/pan/checkpoint-241", "epoch": 3.0, "eval_steps": 500, "global_step": 723, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_bleu": 7.8449, "eval_gen_len": 19.9686, "eval_loss": 1.9624056816101074, "eval_runtime": 23.9819, "eval_samples_per_second": 23.893, "eval_steps_per_second": 3.002, "eval_wer": 0.8248, "step": 241 }, { "epoch": 2.0, "eval_bleu": 2.2342, "eval_gen_len": 17.2356, "eval_loss": 2.161780834197998, "eval_runtime": 23.892, "eval_samples_per_second": 23.983, "eval_steps_per_second": 3.014, "eval_wer": 0.8018, "step": 482 }, { "epoch": 2.074688796680498, "grad_norm": 1.437437653541565, "learning_rate": 4.482365145228216e-05, "loss": 1.8496, "step": 500 }, { "epoch": 3.0, "eval_bleu": 3.4713, "eval_gen_len": 19.5602, "eval_loss": 2.269686698913574, "eval_runtime": 23.8354, "eval_samples_per_second": 24.04, "eval_steps_per_second": 3.021, "eval_wer": 0.7771, "step": 723 }, { "epoch": 3.0, "step": 723, "total_flos": 1410808919629824.0, "train_loss": 1.330777443955713, "train_runtime": 254.2669, "train_samples_per_second": 151.258, "train_steps_per_second": 18.956 } ], "logging_steps": 500, "max_steps": 4820, "num_input_tokens_seen": 0, "num_train_epochs": 20, "save_steps": 50000.0, "stateful_callbacks": { "EarlyStoppingCallback": { "args": { "early_stopping_patience": 2, "early_stopping_threshold": 0.0 }, "attributes": { "early_stopping_patience_counter": 2 } }, "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 1410808919629824.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }