| { | |
| "best_global_step": 2000, | |
| "best_metric": 5.400969970117083, | |
| "best_model_checkpoint": "./SALAMA_NEW/checkpoint-2000", | |
| "epoch": 1.1983223487118035, | |
| "eval_steps": 2000, | |
| "global_step": 2000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2995805871779509, | |
| "grad_norm": 8.785015106201172, | |
| "learning_rate": 9.186875891583453e-06, | |
| "loss": 0.1615, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5991611743559018, | |
| "grad_norm": 4.029974460601807, | |
| "learning_rate": 8.167923374770737e-06, | |
| "loss": 0.1535, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8987417615338527, | |
| "grad_norm": 3.3210387229919434, | |
| "learning_rate": 7.14897085795802e-06, | |
| "loss": 0.1406, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.1983223487118035, | |
| "grad_norm": 2.3362109661102295, | |
| "learning_rate": 6.130018341145303e-06, | |
| "loss": 0.0821, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.1983223487118035, | |
| "eval_loss": 0.0653982162475586, | |
| "eval_runtime": 4767.7684, | |
| "eval_samples_per_second": 2.8, | |
| "eval_steps_per_second": 0.35, | |
| "eval_wer": 5.400969970117083, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5007, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 2000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.846715693727744e+19, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |