checkpoint-1000/trainer_state.json

{
 "best_metric": 1.1078580617904663,
 "best_model_checkpoint": "/kaggle/output/checkpoint-1000",
 "epoch": 0.04074315514993481,
 "eval_steps": 1000,
 "global_step": 1000,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 0.0,
 "learning_rate": 2.7777777777777777e-11,
 "loss": 1.264,
 "step": 1
 },
 {
 "epoch": 0.04,
 "learning_rate": 2.7750000000000004e-08,
 "loss": 1.1723,
 "step": 1000
 },
 {
 "epoch": 0.04,
 "eval_accuracy": 0.3253493013972056,
 "eval_loss": 1.1078580617904663,
 "eval_runtime": 75.5864,
 "eval_samples_per_second": 66.282,
 "eval_steps_per_second": 8.295,
 "step": 1000
 }
 ],
 "logging_steps": 1000,
 "max_steps": 10000000,
 "num_train_epochs": 408,
 "save_steps": 1000,
 "total_flos": 2090355130368000.0,
 "trial_name": null,
 "trial_params": null
}