trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.895193977996526,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.29,
      "learning_rate": 1.8093032233159624e-05,
      "loss": 5.93,
      "step": 500
    },
    {
      "epoch": 0.58,
      "learning_rate": 1.617062343176993e-05,
      "loss": 5.1165,
      "step": 1000
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.4240494113105579e-05,
      "loss": 4.748,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_loss": 4.125702381134033,
      "eval_runtime": 102.7952,
      "eval_samples_per_second": 299.518,
      "eval_steps_per_second": 37.443,
      "step": 1727
    },
    {
      "epoch": 1.16,
      "learning_rate": 1.2314225053078557e-05,
      "loss": 4.5216,
      "step": 2000
    },
    {
      "epoch": 1.45,
      "learning_rate": 1.0384095734414205e-05,
      "loss": 4.3356,
      "step": 2500
    },
    {
      "epoch": 1.74,
      "learning_rate": 8.457826674387185e-06,
      "loss": 4.1988,
      "step": 3000
    },
    {
      "epoch": 2.0,
      "eval_loss": 3.5067617893218994,
      "eval_runtime": 103.1393,
      "eval_samples_per_second": 298.519,
      "eval_steps_per_second": 37.318,
      "step": 3454
    },
    {
      "epoch": 2.03,
      "learning_rate": 6.527697355722834e-06,
      "loss": 4.0849,
      "step": 3500
    },
    {
      "epoch": 2.32,
      "learning_rate": 4.597568037058483e-06,
      "loss": 3.9975,
      "step": 4000
    },
    {
      "epoch": 2.61,
      "learning_rate": 2.6674387183941326e-06,
      "loss": 3.9402,
      "step": 4500
    },
    {
      "epoch": 2.9,
      "learning_rate": 7.37309399729782e-07,
      "loss": 3.8973,
      "step": 5000
    }
  ],
  "max_steps": 5181,
  "num_train_epochs": 3,
  "total_flos": 2.74331103920128e+16,
  "trial_name": null,
  "trial_params": null
}