| { | |
| "best_metric": 1.2666095495224, | |
| "best_model_checkpoint": "/data/tianxing/PycharmProjects/Transformers/examples/conversational/few_shot_intent/file_dir/serialization_dir/checkpoint-6000", | |
| "epoch": 0.17463054724847743, | |
| "eval_steps": 1000, | |
| "global_step": 6000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 0.6197968125343323, | |
| "learning_rate": 0.0002, | |
| "loss": 1.7096, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "eval_loss": 1.4217256307601929, | |
| "eval_runtime": 2455.1897, | |
| "eval_samples_per_second": 118.455, | |
| "eval_steps_per_second": 14.807, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 0.3991679549217224, | |
| "learning_rate": 0.00019400443671682956, | |
| "loss": 1.2426, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "eval_loss": 1.3142809867858887, | |
| "eval_runtime": 2454.5071, | |
| "eval_samples_per_second": 118.488, | |
| "eval_steps_per_second": 14.811, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 0.3880527913570404, | |
| "learning_rate": 0.0001880088734336591, | |
| "loss": 1.1529, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "eval_loss": 1.2872347831726074, | |
| "eval_runtime": 2455.0423, | |
| "eval_samples_per_second": 118.462, | |
| "eval_steps_per_second": 14.808, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 0.3351105749607086, | |
| "learning_rate": 0.00018201331015048865, | |
| "loss": 1.0968, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "eval_loss": 1.2708055973052979, | |
| "eval_runtime": 2454.672, | |
| "eval_samples_per_second": 118.48, | |
| "eval_steps_per_second": 14.81, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 0.42601627111434937, | |
| "learning_rate": 0.0001760177468673182, | |
| "loss": 1.0568, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "eval_loss": 1.2738263607025146, | |
| "eval_runtime": 2454.9426, | |
| "eval_samples_per_second": 118.467, | |
| "eval_steps_per_second": 14.808, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 0.3943077027797699, | |
| "learning_rate": 0.00017002218358414775, | |
| "loss": 1.0274, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "eval_loss": 1.2666095495224, | |
| "eval_runtime": 2453.7183, | |
| "eval_samples_per_second": 118.526, | |
| "eval_steps_per_second": 14.816, | |
| "step": 6000 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 34358, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "total_flos": 1.9766892483628237e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |