checkpoint-1000/trainer_state.json

{
 "best_metric": 1.1557353734970093,
 "best_model_checkpoint": "/kaggle/output/checkpoint-1000",
 "epoch": 16.0,
 "eval_steps": 1000,
 "global_step": 1000,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 0.02,
 "learning_rate": 2.7777777777777777e-11,
 "loss": 1.2395,
 "step": 1
 },
 {
 "epoch": 16.0,
 "learning_rate": 2.7638888888888893e-08,
 "loss": 1.1422,
 "step": 1000
 },
 {
 "epoch": 16.0,
 "eval_accuracy": 0.33073852295409184,
 "eval_loss": 1.1557353734970093,
 "eval_runtime": 54.1805,
 "eval_samples_per_second": 92.469,
 "eval_steps_per_second": 11.572,
 "step": 1000
 }
 ],
 "logging_steps": 1000,
 "max_steps": 10000000,
 "num_train_epochs": 161291,
 "save_steps": 1000,
 "total_flos": 2090355130368000.0,
 "trial_name": null,
 "trial_params": null
}