checkpoint-500/trainer_state.json

{
  "best_metric": 0.42523157596588135,
  "best_model_checkpoint": "mikhail-panzo/fil_b128_le4_s8000/checkpoint-500",
  "epoch": 44.19889502762431,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.419889502762431,
      "grad_norm": 2.1200406551361084,
      "learning_rate": 2.5e-06,
      "loss": 0.77,
      "step": 50
    },
    {
      "epoch": 8.839779005524862,
      "grad_norm": 0.8919622302055359,
      "learning_rate": 5e-06,
      "loss": 0.695,
      "step": 100
    },
    {
      "epoch": 13.259668508287293,
      "grad_norm": 0.9864558577537537,
      "learning_rate": 7.5e-06,
      "loss": 0.6278,
      "step": 150
    },
    {
      "epoch": 17.679558011049725,
      "grad_norm": 0.9195986986160278,
      "learning_rate": 1e-05,
      "loss": 0.5431,
      "step": 200
    },
    {
      "epoch": 22.099447513812155,
      "grad_norm": 1.0175281763076782,
      "learning_rate": 1.25e-05,
      "loss": 0.5155,
      "step": 250
    },
    {
      "epoch": 26.519337016574585,
      "grad_norm": 1.0303566455841064,
      "learning_rate": 1.5e-05,
      "loss": 0.502,
      "step": 300
    },
    {
      "epoch": 30.939226519337016,
      "grad_norm": 1.005897879600525,
      "learning_rate": 1.75e-05,
      "loss": 0.4889,
      "step": 350
    },
    {
      "epoch": 35.35911602209945,
      "grad_norm": 1.0381360054016113,
      "learning_rate": 2e-05,
      "loss": 0.477,
      "step": 400
    },
    {
      "epoch": 39.77900552486188,
      "grad_norm": 1.2148911952972412,
      "learning_rate": 2.25e-05,
      "loss": 0.4711,
      "step": 450
    },
    {
      "epoch": 44.19889502762431,
      "grad_norm": 1.5404785871505737,
      "learning_rate": 2.5e-05,
      "loss": 0.463,
      "step": 500
    },
    {
      "epoch": 44.19889502762431,
      "eval_loss": 0.42523157596588135,
      "eval_runtime": 15.6706,
      "eval_samples_per_second": 10.274,
      "eval_steps_per_second": 1.34,
      "step": 500
    }
  ],
  "logging_steps": 50,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 728,
  "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": 1.1322540086397408e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}