trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.96,
  "eval_steps": 100,
  "global_step": 6,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 555.18984375,
      "epoch": 1.6400000000000001,
      "grad_norm": 0.9047526121139526,
      "kl": 22.304883193969726,
      "learning_rate": 1.9098300562505264e-07,
      "loss": 0.9364,
      "reward": 0.47890625,
      "reward_std": 0.3222931930795312,
      "rewards/accuracy_reward": 0.07734375,
      "rewards/format_reward": 0.009375,
      "rewards/tag_count_reward": 0.3921875,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 546.359375,
      "epoch": 1.96,
      "kl": 0.04564642906188965,
      "reward": 0.5244140625,
      "reward_std": 0.2628240427002311,
      "rewards/accuracy_reward": 0.01953125,
      "rewards/format_reward": 0.01171875,
      "rewards/tag_count_reward": 0.4931640625,
      "step": 6,
      "total_flos": 0.0,
      "train_loss": 0.7375955581665039,
      "train_runtime": 1563.9615,
      "train_samples_per_second": 0.128,
      "train_steps_per_second": 0.004
    }
  ],
  "logging_steps": 5,
  "max_steps": 6,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}