trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 12.0,
  "eval_steps": 20,
  "global_step": 3024,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.6863582730293274,
      "eval_runtime": 81.6224,
      "eval_samples_per_second": 2.646,
      "eval_steps_per_second": 0.331,
      "step": 252
    },
    {
      "epoch": 1.9841269841269842,
      "grad_norm": 5.002901077270508,
      "learning_rate": 2.256206554121152e-05,
      "loss": 0.887,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.6000874638557434,
      "eval_runtime": 81.7972,
      "eval_samples_per_second": 2.641,
      "eval_steps_per_second": 0.33,
      "step": 504
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.590830385684967,
      "eval_runtime": 81.599,
      "eval_samples_per_second": 2.647,
      "eval_steps_per_second": 0.331,
      "step": 756
    },
    {
      "epoch": 3.9682539682539684,
      "grad_norm": 6.367453575134277,
      "learning_rate": 2.0079443892750745e-05,
      "loss": 0.3601,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.609752893447876,
      "eval_runtime": 81.6199,
      "eval_samples_per_second": 2.646,
      "eval_steps_per_second": 0.331,
      "step": 1008
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.6210949420928955,
      "eval_runtime": 81.5698,
      "eval_samples_per_second": 2.648,
      "eval_steps_per_second": 0.331,
      "step": 1260
    },
    {
      "epoch": 5.9523809523809526,
      "grad_norm": 5.3468403816223145,
      "learning_rate": 1.759682224428997e-05,
      "loss": 0.2094,
      "step": 1500
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.643219530582428,
      "eval_runtime": 81.5714,
      "eval_samples_per_second": 2.648,
      "eval_steps_per_second": 0.331,
      "step": 1512
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.6334556937217712,
      "eval_runtime": 81.7437,
      "eval_samples_per_second": 2.642,
      "eval_steps_per_second": 0.33,
      "step": 1764
    },
    {
      "epoch": 7.936507936507937,
      "grad_norm": 1.6176210641860962,
      "learning_rate": 1.512909632571996e-05,
      "loss": 0.1354,
      "step": 2000
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.663597822189331,
      "eval_runtime": 81.7945,
      "eval_samples_per_second": 2.641,
      "eval_steps_per_second": 0.33,
      "step": 2016
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.6830390691757202,
      "eval_runtime": 81.6219,
      "eval_samples_per_second": 2.646,
      "eval_steps_per_second": 0.331,
      "step": 2268
    },
    {
      "epoch": 9.920634920634921,
      "grad_norm": 3.5583207607269287,
      "learning_rate": 1.2646474677259188e-05,
      "loss": 0.1201,
      "step": 2500
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.6750147938728333,
      "eval_runtime": 81.6464,
      "eval_samples_per_second": 2.646,
      "eval_steps_per_second": 0.331,
      "step": 2520
    },
    {
      "epoch": 11.0,
      "eval_loss": 0.6730176210403442,
      "eval_runtime": 81.6583,
      "eval_samples_per_second": 2.645,
      "eval_steps_per_second": 0.331,
      "step": 2772
    },
    {
      "epoch": 11.904761904761905,
      "grad_norm": 0.9115921258926392,
      "learning_rate": 1.0163853028798411e-05,
      "loss": 0.1093,
      "step": 3000
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.702785074710846,
      "eval_runtime": 81.6367,
      "eval_samples_per_second": 2.646,
      "eval_steps_per_second": 0.331,
      "step": 3024
    }
  ],
  "logging_steps": 500,
  "max_steps": 5040,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 1,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.4079475431191347e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}