trainer_state.json

{
  "best_metric": 1.276187777519226,
  "best_model_checkpoint": "shawgpt-ft-trial-5/checkpoint-15",
  "epoch": 8.64,
  "eval_steps": 500,
  "global_step": 27,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.96,
      "grad_norm": 1.3013744354248047,
      "learning_rate": 0.0008885909996223793,
      "loss": 3.779,
      "step": 3
    },
    {
      "epoch": 0.96,
      "eval_loss": 2.497284412384033,
      "eval_runtime": 4.481,
      "eval_samples_per_second": 2.008,
      "eval_steps_per_second": 1.116,
      "step": 3
    },
    {
      "epoch": 1.92,
      "grad_norm": 2.2233452796936035,
      "learning_rate": 0.000777517124669582,
      "loss": 2.1425,
      "step": 6
    },
    {
      "epoch": 1.92,
      "eval_loss": 1.594606876373291,
      "eval_runtime": 4.4761,
      "eval_samples_per_second": 2.011,
      "eval_steps_per_second": 1.117,
      "step": 6
    },
    {
      "epoch": 2.88,
      "grad_norm": 0.5042614936828613,
      "learning_rate": 0.0006664432497167844,
      "loss": 1.3668,
      "step": 9
    },
    {
      "epoch": 2.88,
      "eval_loss": 1.32815682888031,
      "eval_runtime": 4.478,
      "eval_samples_per_second": 2.01,
      "eval_steps_per_second": 1.117,
      "step": 9
    },
    {
      "epoch": 3.84,
      "grad_norm": 0.39059099555015564,
      "learning_rate": 0.0005553693747639871,
      "loss": 1.1703,
      "step": 12
    },
    {
      "epoch": 3.84,
      "eval_loss": 1.2933988571166992,
      "eval_runtime": 4.4781,
      "eval_samples_per_second": 2.01,
      "eval_steps_per_second": 1.117,
      "step": 12
    },
    {
      "epoch": 4.8,
      "grad_norm": 0.35272151231765747,
      "learning_rate": 0.00044429549981118966,
      "loss": 1.0645,
      "step": 15
    },
    {
      "epoch": 4.8,
      "eval_loss": 1.276187777519226,
      "eval_runtime": 4.4763,
      "eval_samples_per_second": 2.011,
      "eval_steps_per_second": 1.117,
      "step": 15
    },
    {
      "epoch": 5.76,
      "grad_norm": 0.3423387110233307,
      "learning_rate": 0.0003332216248583922,
      "loss": 1.007,
      "step": 18
    },
    {
      "epoch": 5.76,
      "eval_loss": 1.2889482975006104,
      "eval_runtime": 4.4767,
      "eval_samples_per_second": 2.01,
      "eval_steps_per_second": 1.117,
      "step": 18
    },
    {
      "epoch": 6.72,
      "grad_norm": 0.359205424785614,
      "learning_rate": 0.00022214774990559483,
      "loss": 0.9542,
      "step": 21
    },
    {
      "epoch": 6.72,
      "eval_loss": 1.3014657497406006,
      "eval_runtime": 4.4819,
      "eval_samples_per_second": 2.008,
      "eval_steps_per_second": 1.116,
      "step": 21
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.38613542914390564,
      "learning_rate": 7.404924996853161e-05,
      "loss": 0.6906,
      "step": 25
    },
    {
      "epoch": 8.0,
      "eval_loss": 1.3172171115875244,
      "eval_runtime": 4.4776,
      "eval_samples_per_second": 2.01,
      "eval_steps_per_second": 1.117,
      "step": 25
    },
    {
      "epoch": 8.64,
      "grad_norm": 0.3728366494178772,
      "learning_rate": 0.0,
      "loss": 0.8112,
      "step": 27
    },
    {
      "epoch": 8.64,
      "eval_loss": 1.3243048191070557,
      "eval_runtime": 4.4699,
      "eval_samples_per_second": 2.013,
      "eval_steps_per_second": 1.119,
      "step": 27
    },
    {
      "epoch": 8.64,
      "step": 27,
      "total_flos": 81057677524992.0,
      "train_loss": 1.438443453223617,
      "train_runtime": 476.4437,
      "train_samples_per_second": 0.944,
      "train_steps_per_second": 0.057
    }
  ],
  "logging_steps": 500,
  "max_steps": 27,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 81057677524992.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}