| { | |
| "best_metric": 0.002982336562126875, | |
| "best_model_checkpoint": "t5/checkpoint-25566", | |
| "epoch": 8.0, | |
| "eval_steps": 500, | |
| "global_step": 102264, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.09756029397249222, | |
| "learning_rate": 1.9600000000000002e-05, | |
| "loss": 0.0059, | |
| "step": 12783 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9804666666666667, | |
| "eval_loss": 0.0031315775122493505, | |
| "eval_runtime": 472.802, | |
| "eval_samples_per_second": 63.452, | |
| "eval_steps_per_second": 0.497, | |
| "step": 12783 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.0030228213872760534, | |
| "learning_rate": 1.9200000000000003e-05, | |
| "loss": 0.0022, | |
| "step": 25566 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9818666666666667, | |
| "eval_loss": 0.002982336562126875, | |
| "eval_runtime": 474.0664, | |
| "eval_samples_per_second": 63.282, | |
| "eval_steps_per_second": 0.496, | |
| "step": 25566 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.00019585879635997117, | |
| "learning_rate": 1.88e-05, | |
| "loss": 0.0012, | |
| "step": 38349 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9825, | |
| "eval_loss": 0.0032824024092406034, | |
| "eval_runtime": 478.2854, | |
| "eval_samples_per_second": 62.724, | |
| "eval_steps_per_second": 0.491, | |
| "step": 38349 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.004981159698218107, | |
| "learning_rate": 1.8400000000000003e-05, | |
| "loss": 0.0007, | |
| "step": 51132 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9826, | |
| "eval_loss": 0.003532745875418186, | |
| "eval_runtime": 469.3109, | |
| "eval_samples_per_second": 63.924, | |
| "eval_steps_per_second": 0.501, | |
| "step": 51132 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.002326933667063713, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.0005, | |
| "step": 63915 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9821666666666666, | |
| "eval_loss": 0.003947438672184944, | |
| "eval_runtime": 471.3239, | |
| "eval_samples_per_second": 63.65, | |
| "eval_steps_per_second": 0.499, | |
| "step": 63915 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.034791357815265656, | |
| "learning_rate": 1.76e-05, | |
| "loss": 0.0004, | |
| "step": 76698 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9815666666666667, | |
| "eval_loss": 0.0041623483411967754, | |
| "eval_runtime": 470.7737, | |
| "eval_samples_per_second": 63.725, | |
| "eval_steps_per_second": 0.499, | |
| "step": 76698 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 9.868334018392488e-05, | |
| "learning_rate": 1.72e-05, | |
| "loss": 0.0003, | |
| "step": 89481 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9822666666666666, | |
| "eval_loss": 0.004203655291348696, | |
| "eval_runtime": 471.0374, | |
| "eval_samples_per_second": 63.689, | |
| "eval_steps_per_second": 0.499, | |
| "step": 89481 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.002833570586517453, | |
| "learning_rate": 1.6800000000000002e-05, | |
| "loss": 0.0003, | |
| "step": 102264 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9826, | |
| "eval_loss": 0.0045397114008665085, | |
| "eval_runtime": 469.6081, | |
| "eval_samples_per_second": 63.883, | |
| "eval_steps_per_second": 0.5, | |
| "step": 102264 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 639150, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 500, | |
| "total_flos": 5.937159320662856e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |