checkpoint-50/trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.5989590287208557,
  "best_model_checkpoint": "llm/adapters/Llama-3.1-Minitron_cr_chaptered/checkpoint-20",
  "epoch": 50.0,
  "eval_steps": 10,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 10.0,
      "grad_norm": 0.3206774890422821,
      "learning_rate": 0.00027299999999999997,
      "loss": 1.1999,
      "step": 10
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.6795049905776978,
      "eval_runtime": 3.3507,
      "eval_samples_per_second": 8.655,
      "eval_steps_per_second": 8.655,
      "step": 10
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.12689660489559174,
      "learning_rate": 0.000243,
      "loss": 0.3747,
      "step": 20
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.5989590287208557,
      "eval_runtime": 3.3773,
      "eval_samples_per_second": 8.587,
      "eval_steps_per_second": 8.587,
      "step": 20
    },
    {
      "epoch": 30.0,
      "grad_norm": 0.08760394155979156,
      "learning_rate": 0.00021299999999999997,
      "loss": 0.2734,
      "step": 30
    },
    {
      "epoch": 30.0,
      "eval_loss": 0.6167947053909302,
      "eval_runtime": 3.3875,
      "eval_samples_per_second": 8.561,
      "eval_steps_per_second": 8.561,
      "step": 30
    },
    {
      "epoch": 40.0,
      "grad_norm": 0.03870585188269615,
      "learning_rate": 0.00018299999999999998,
      "loss": 0.2443,
      "step": 40
    },
    {
      "epoch": 40.0,
      "eval_loss": 0.6276828050613403,
      "eval_runtime": 3.3822,
      "eval_samples_per_second": 8.574,
      "eval_steps_per_second": 8.574,
      "step": 40
    },
    {
      "epoch": 50.0,
      "grad_norm": 0.04375430941581726,
      "learning_rate": 0.00015299999999999998,
      "loss": 0.2363,
      "step": 50
    },
    {
      "epoch": 50.0,
      "eval_loss": 0.6464866399765015,
      "eval_runtime": 3.3731,
      "eval_samples_per_second": 8.598,
      "eval_steps_per_second": 8.598,
      "step": 50
    }
  ],
  "logging_steps": 10,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 10,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.59015595763712e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}