trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 9.67741935483871,
  "eval_steps": 40,
  "global_step": 150,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.6451612903225806,
      "grad_norm": 0.8625742793083191,
      "learning_rate": 4e-05,
      "loss": 0.7057,
      "step": 10
    },
    {
      "epoch": 1.2903225806451613,
      "grad_norm": 0.7124119400978088,
      "learning_rate": 8e-05,
      "loss": 0.6741,
      "step": 20
    },
    {
      "epoch": 1.935483870967742,
      "grad_norm": 0.7693650126457214,
      "learning_rate": 0.00012,
      "loss": 0.5941,
      "step": 30
    },
    {
      "epoch": 2.5806451612903225,
      "grad_norm": 0.6899323463439941,
      "learning_rate": 0.00016,
      "loss": 0.547,
      "step": 40
    },
    {
      "epoch": 2.5806451612903225,
      "eval_loss": 0.48405569791793823,
      "eval_runtime": 2.1967,
      "eval_samples_per_second": 0.455,
      "eval_steps_per_second": 0.455,
      "step": 40
    },
    {
      "epoch": 3.225806451612903,
      "grad_norm": 0.564319908618927,
      "learning_rate": 0.0002,
      "loss": 0.5046,
      "step": 50
    },
    {
      "epoch": 3.870967741935484,
      "grad_norm": 0.6034739017486572,
      "learning_rate": 0.00024,
      "loss": 0.4842,
      "step": 60
    },
    {
      "epoch": 4.516129032258064,
      "grad_norm": 0.6102902293205261,
      "learning_rate": 0.00028,
      "loss": 0.4531,
      "step": 70
    },
    {
      "epoch": 5.161290322580645,
      "grad_norm": 0.7875820398330688,
      "learning_rate": 0.00032,
      "loss": 0.4389,
      "step": 80
    },
    {
      "epoch": 5.161290322580645,
      "eval_loss": 0.4346591830253601,
      "eval_runtime": 2.1674,
      "eval_samples_per_second": 0.461,
      "eval_steps_per_second": 0.461,
      "step": 80
    },
    {
      "epoch": 5.806451612903226,
      "grad_norm": 0.8663121461868286,
      "learning_rate": 0.00036,
      "loss": 0.4177,
      "step": 90
    },
    {
      "epoch": 6.451612903225806,
      "grad_norm": 0.9721035957336426,
      "learning_rate": 0.0004,
      "loss": 0.4046,
      "step": 100
    },
    {
      "epoch": 7.096774193548387,
      "grad_norm": 0.9510289430618286,
      "learning_rate": 0.00032,
      "loss": 0.3788,
      "step": 110
    },
    {
      "epoch": 7.741935483870968,
      "grad_norm": 1.0126413106918335,
      "learning_rate": 0.00024,
      "loss": 0.3673,
      "step": 120
    },
    {
      "epoch": 7.741935483870968,
      "eval_loss": 0.4241188168525696,
      "eval_runtime": 2.195,
      "eval_samples_per_second": 0.456,
      "eval_steps_per_second": 0.456,
      "step": 120
    },
    {
      "epoch": 8.387096774193548,
      "grad_norm": 1.116309642791748,
      "learning_rate": 0.00016,
      "loss": 0.3449,
      "step": 130
    },
    {
      "epoch": 9.03225806451613,
      "grad_norm": 0.9966815710067749,
      "learning_rate": 8e-05,
      "loss": 0.3338,
      "step": 140
    },
    {
      "epoch": 9.67741935483871,
      "grad_norm": 1.0189791917800903,
      "learning_rate": 0.0,
      "loss": 0.3202,
      "step": 150
    }
  ],
  "logging_steps": 10,
  "max_steps": 150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 40,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.946319309952778e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}