checkpoint-372/trainer_state.json

{
  "best_global_step": null,
  "best_metric": Infinity,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 372,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.40547389761784086,
      "grad_norm": 0.3029731810092926,
      "learning_rate": 0.00019341317365269463,
      "loss": 0.7336131286621094,
      "step": 50
    },
    {
      "epoch": 0.8109477952356817,
      "grad_norm": 0.6691709756851196,
      "learning_rate": 0.00016347305389221556,
      "loss": 0.5254125595092773,
      "step": 100
    },
    {
      "epoch": 1.0,
      "eval_loss": NaN,
      "eval_runtime": 520.4087,
      "eval_samples_per_second": 0.575,
      "eval_steps_per_second": 0.575,
      "step": 124
    },
    {
      "epoch": 1.2108464267612773,
      "grad_norm": 0.6433291435241699,
      "learning_rate": 0.00013353293413173654,
      "loss": 0.5045146179199219,
      "step": 150
    },
    {
      "epoch": 1.6163203243791182,
      "grad_norm": 0.4409848749637604,
      "learning_rate": 0.0001035928143712575,
      "loss": 0.4645370101928711,
      "step": 200
    },
    {
      "epoch": 2.0,
      "eval_loss": NaN,
      "eval_runtime": 521.254,
      "eval_samples_per_second": 0.574,
      "eval_steps_per_second": 0.574,
      "step": 248
    },
    {
      "epoch": 2.0162189559047135,
      "grad_norm": 0.46470925211906433,
      "learning_rate": 7.365269461077845e-05,
      "loss": 0.4020792770385742,
      "step": 250
    },
    {
      "epoch": 2.4216928535225546,
      "grad_norm": 0.36329224705696106,
      "learning_rate": 4.3712574850299406e-05,
      "loss": 0.4243511199951172,
      "step": 300
    },
    {
      "epoch": 2.8271667511403953,
      "grad_norm": 0.3943828046321869,
      "learning_rate": 1.377245508982036e-05,
      "loss": 0.40013973236083983,
      "step": 350
    },
    {
      "epoch": 3.0,
      "eval_loss": NaN,
      "eval_runtime": 523.2415,
      "eval_samples_per_second": 0.571,
      "eval_steps_per_second": 0.571,
      "step": 372
    }
  ],
  "logging_steps": 50,
  "max_steps": 372,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 1.3147963552353485e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}