checkpoint-3000/trainer_state.json

{
 "best_metric": 1.1441665887832642,
 "best_model_checkpoint": "/kaggle/output/checkpoint-3000",
 "epoch": 48.0,
 "eval_steps": 1000,
 "global_step": 3000,
 "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
 },
 {
 "epoch": 32.0,
 "learning_rate": 5.541666666666667e-08,
 "loss": 1.1401,
 "step": 2000
 },
 {
 "epoch": 32.0,
 "eval_accuracy": 0.3311377245508982,
 "eval_loss": 1.1507551670074463,
 "eval_runtime": 53.98,
 "eval_samples_per_second": 92.812,
 "eval_steps_per_second": 11.615,
 "step": 2000
 },
 {
 "epoch": 48.0,
 "learning_rate": 8.316666666666666e-08,
 "loss": 1.1287,
 "step": 3000
 },
 {
 "epoch": 48.0,
 "eval_accuracy": 0.3321357285429142,
 "eval_loss": 1.1441665887832642,
 "eval_runtime": 53.9873,
 "eval_samples_per_second": 92.8,
 "eval_steps_per_second": 11.614,
 "step": 3000
 }
 ],
 "logging_steps": 1000,
 "max_steps": 10000000,
 "num_train_epochs": 161291,
 "save_steps": 1000,
 "total_flos": 6271065391104000.0,
 "trial_name": null,
 "trial_params": null
}