checkpoint-4000/trainer_state.json

{
 "best_metric": 1.1194497346878052,
 "best_model_checkpoint": "/kaggle/output/checkpoint-4000",
 "epoch": 0.16297262059973924,
 "eval_steps": 1000,
 "global_step": 4000,
 "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": 1.029,
 "step": 1
 },
 {
 "epoch": 0.04,
 "learning_rate": 2.7638888888888893e-08,
 "loss": 1.181,
 "step": 1000
 },
 {
 "epoch": 0.04,
 "eval_accuracy": 0.3273453093812375,
 "eval_loss": 1.1529844999313354,
 "eval_runtime": 54.2837,
 "eval_samples_per_second": 92.293,
 "eval_steps_per_second": 11.55,
 "step": 1000
 },
 {
 "epoch": 0.08,
 "learning_rate": 5.541666666666667e-08,
 "loss": 1.1527,
 "step": 2000
 },
 {
 "epoch": 0.08,
 "eval_accuracy": 0.33013972055888224,
 "eval_loss": 1.1351025104522705,
 "eval_runtime": 54.0518,
 "eval_samples_per_second": 92.689,
 "eval_steps_per_second": 11.6,
 "step": 2000
 },
 {
 "epoch": 0.12,
 "learning_rate": 8.316666666666666e-08,
 "loss": 1.142,
 "step": 3000
 },
 {
 "epoch": 0.12,
 "eval_accuracy": 0.3317365269461078,
 "eval_loss": 1.127414345741272,
 "eval_runtime": 54.0871,
 "eval_samples_per_second": 92.628,
 "eval_steps_per_second": 11.592,
 "step": 3000
 },
 {
 "epoch": 0.16,
 "learning_rate": 1.1091666666666668e-07,
 "loss": 1.1371,
 "step": 4000
 },
 {
 "epoch": 0.16,
 "eval_accuracy": 0.331936127744511,
 "eval_loss": 1.1194497346878052,
 "eval_runtime": 54.3907,
 "eval_samples_per_second": 92.111,
 "eval_steps_per_second": 11.528,
 "step": 4000
 }
 ],
 "logging_steps": 1000,
 "max_steps": 10000000,
 "num_train_epochs": 408,
 "save_steps": 1000,
 "total_flos": 8361420521472000.0,
 "trial_name": null,
 "trial_params": null
}