trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9924812030075187,
  "eval_steps": 500,
  "global_step": 66,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08,
      "learning_rate": 4.929528920808854e-05,
      "loss": 3.4468,
      "step": 5
    },
    {
      "epoch": 0.15,
      "learning_rate": 4.722088621637309e-05,
      "loss": 3.5567,
      "step": 10
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.389373935885646e-05,
      "loss": 2.5191,
      "step": 15
    },
    {
      "epoch": 0.3,
      "learning_rate": 3.9501422739279956e-05,
      "loss": 2.0085,
      "step": 20
    },
    {
      "epoch": 0.38,
      "learning_rate": 3.4291561391508185e-05,
      "loss": 1.475,
      "step": 25
    },
    {
      "epoch": 0.45,
      "learning_rate": 2.8557870956832132e-05,
      "loss": 1.462,
      "step": 30
    },
    {
      "epoch": 0.53,
      "learning_rate": 2.2623598917395438e-05,
      "loss": 1.1812,
      "step": 35
    },
    {
      "epoch": 0.6,
      "learning_rate": 1.682330091706446e-05,
      "loss": 1.2592,
      "step": 40
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.148397956361007e-05,
      "loss": 1.0798,
      "step": 45
    },
    {
      "epoch": 0.75,
      "learning_rate": 6.906649047373246e-06,
      "loss": 1.0623,
      "step": 50
    },
    {
      "epoch": 0.83,
      "learning_rate": 3.3493649053890326e-06,
      "loss": 1.346,
      "step": 55
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.0126756596375686e-06,
      "loss": 1.1372,
      "step": 60
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.831652042480093e-08,
      "loss": 1.047,
      "step": 65
    },
    {
      "epoch": 0.99,
      "step": 66,
      "total_flos": 4369517458489344.0,
      "train_loss": 1.7228616259314797,
      "train_runtime": 47.9316,
      "train_samples_per_second": 22.157,
      "train_steps_per_second": 1.377
    }
  ],
  "logging_steps": 5,
  "max_steps": 66,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 4369517458489344.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}