trainer_state.json

{
 "best_metric": null,
 "best_model_checkpoint": null,
 "epoch": 80.0,
 "eval_steps": 100,
 "global_step": 400,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 20.0,
 "grad_norm": 0.009852898307144642,
 "learning_rate": 0.00015000000000000001,
 "loss": 0.1022,
 "step": 100
 },
 {
 "epoch": 20.0,
 "eval_accuracy": 1.0,
 "eval_loss": 0.0008228308288380504,
 "eval_runtime": 0.0347,
 "eval_samples_per_second": 519.229,
 "eval_steps_per_second": 57.692,
 "step": 100
 },
 {
 "epoch": 40.0,
 "grad_norm": 0.008125479333102703,
 "learning_rate": 0.0001,
 "loss": 0.006,
 "step": 200
 },
 {
 "epoch": 40.0,
 "eval_accuracy": 1.0,
 "eval_loss": 0.0006637517362833023,
 "eval_runtime": 0.028,
 "eval_samples_per_second": 642.488,
 "eval_steps_per_second": 71.388,
 "step": 200
 },
 {
 "epoch": 60.0,
 "grad_norm": 0.00710184033960104,
 "learning_rate": 5e-05,
 "loss": 0.0223,
 "step": 300
 },
 {
 "epoch": 60.0,
 "eval_accuracy": 1.0,
 "eval_loss": 0.0019779244903475046,
 "eval_runtime": 0.0323,
 "eval_samples_per_second": 557.646,
 "eval_steps_per_second": 61.961,
 "step": 300
 },
 {
 "epoch": 80.0,
 "grad_norm": 0.00847003422677517,
 "learning_rate": 0.0,
 "loss": 0.0058,
 "step": 400
 },
 {
 "epoch": 80.0,
 "eval_accuracy": 1.0,
 "eval_loss": 0.0008943404536694288,
 "eval_runtime": 0.032,
 "eval_samples_per_second": 562.516,
 "eval_steps_per_second": 62.502,
 "step": 400
 },
 {
 "epoch": 80.0,
 "step": 400,
 "total_flos": 66867808427280.0,
 "train_loss": 0.03407109707593918,
 "train_runtime": 41.4151,
 "train_samples_per_second": 154.533,
 "train_steps_per_second": 9.658
 }
 ],
 "logging_steps": 100,
 "max_steps": 400,
 "num_input_tokens_seen": 0,
 "num_train_epochs": 80,
 "save_steps": 100,
 "total_flos": 66867808427280.0,
 "train_batch_size": 16,
 "trial_name": null,
 "trial_params": null
}