checkpoint-500/trainer_state.json

{
  "best_metric": 0.40509888529777527,
  "best_model_checkpoint": "mikhail_panzo/ceb_b128_le3_s4000/checkpoint-500",
  "epoch": 39.603960396039604,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 3.9603960396039604,
      "grad_norm": 1.6000345945358276,
      "learning_rate": 2.5e-05,
      "loss": 0.7102,
      "step": 50
    },
    {
      "epoch": 7.920792079207921,
      "grad_norm": 1.240767478942871,
      "learning_rate": 5e-05,
      "loss": 0.5129,
      "step": 100
    },
    {
      "epoch": 11.881188118811881,
      "grad_norm": 2.545292377471924,
      "learning_rate": 7.5e-05,
      "loss": 0.4701,
      "step": 150
    },
    {
      "epoch": 15.841584158415841,
      "grad_norm": 3.054067611694336,
      "learning_rate": 0.0001,
      "loss": 0.4502,
      "step": 200
    },
    {
      "epoch": 19.801980198019802,
      "grad_norm": 3.884960889816284,
      "learning_rate": 0.000125,
      "loss": 0.4433,
      "step": 250
    },
    {
      "epoch": 23.762376237623762,
      "grad_norm": 2.102992296218872,
      "learning_rate": 0.00015,
      "loss": 0.4361,
      "step": 300
    },
    {
      "epoch": 27.722772277227723,
      "grad_norm": 3.1947975158691406,
      "learning_rate": 0.000175,
      "loss": 0.4335,
      "step": 350
    },
    {
      "epoch": 31.683168316831683,
      "grad_norm": 4.315099716186523,
      "learning_rate": 0.0002,
      "loss": 0.4322,
      "step": 400
    },
    {
      "epoch": 35.64356435643565,
      "grad_norm": 6.909496784210205,
      "learning_rate": 0.00022500000000000002,
      "loss": 0.4346,
      "step": 450
    },
    {
      "epoch": 39.603960396039604,
      "grad_norm": 2.0800621509552,
      "learning_rate": 0.00025,
      "loss": 0.42,
      "step": 500
    },
    {
      "epoch": 39.603960396039604,
      "eval_loss": 0.40509888529777527,
      "eval_runtime": 8.3422,
      "eval_samples_per_second": 21.577,
      "eval_steps_per_second": 2.757,
      "step": 500
    }
  ],
  "logging_steps": 50,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 334,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0803690034289064e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}