trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00291216898345888,
  "eval_steps": 500,
  "global_step": 300,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0004853614972431467,
      "frac_reward_zero_std": 0.52,
      "grad_norm": 6.388103793142363e-05,
      "kl": 0.0009641998598817736,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 0.0,
      "num_tokens": 222000.0,
      "reward": 0.445,
      "reward_std": 0.21117892980575562,
      "rewards/compute_reward/mean": 0.445,
      "rewards/compute_reward/std": 0.211178936958313,
      "step": 50
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0009707229944862934,
      "frac_reward_zero_std": 0.64,
      "grad_norm": 0.0001019326300593093,
      "kl": 0.0010897845961153508,
      "learning_rate": 4.95e-06,
      "loss": 0.0,
      "num_tokens": 444000.0,
      "reward": 0.415,
      "reward_std": 0.15713949322700502,
      "rewards/compute_reward/mean": 0.415,
      "rewards/compute_reward/std": 0.15713950037956237,
      "step": 100
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.00145608449172944,
      "frac_reward_zero_std": 0.48,
      "grad_norm": 0.0002248563978355378,
      "kl": 0.0016197766526602209,
      "learning_rate": 4.963519746208726e-06,
      "loss": 0.0,
      "num_tokens": 666000.0,
      "reward": 0.435,
      "reward_std": 0.23550184726715087,
      "rewards/compute_reward/mean": 0.435,
      "rewards/compute_reward/std": 0.23550185441970825,
      "step": 150
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0019414459889725867,
      "frac_reward_zero_std": 0.6,
      "grad_norm": 0.0003219710197299719,
      "kl": 0.006518472717143595,
      "learning_rate": 4.852201922385564e-06,
      "loss": 0.0,
      "num_tokens": 888000.0,
      "reward": 0.3675,
      "reward_std": 0.1800285929441452,
      "rewards/compute_reward/mean": 0.3675,
      "rewards/compute_reward/std": 0.18002860248088837,
      "step": 200
    },
    {
      "completion_length": 512.0,
      "completions/clipped_ratio": 1.0,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 0.0,
      "completions/mean_length": 512.0,
      "completions/mean_terminated_length": 0.0,
      "completions/min_length": 512.0,
      "completions/min_terminated_length": 0.0,
      "epoch": 0.0024268074862157335,
      "frac_reward_zero_std": 0.64,
      "grad_norm": 0.32203951478004456,
      "kl": 0.007433977099135518,
      "learning_rate": 4.669413633422322e-06,
      "loss": 0.0,
      "num_tokens": 1110000.0,
      "reward": 0.315,
      "reward_std": 0.14839409828186034,
      "rewards/compute_reward/mean": 0.315,
      "rewards/compute_reward/std": 0.14839410066604614,
      "step": 250
    },
    {
      "completion_length": 511.4675,
      "completions/clipped_ratio": 0.9975,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 5.98,
      "completions/mean_length": 511.4675,
      "completions/mean_terminated_length": 5.98,
      "completions/min_length": 507.74,
      "completions/min_terminated_length": 5.98,
      "epoch": 0.00291216898345888,
      "frac_reward_zero_std": 0.52,
      "grad_norm": 0.0001669295015744865,
      "kl": 0.006612616684287786,
      "learning_rate": 4.420708808983809e-06,
      "loss": 0.0,
      "num_tokens": 1331787.0,
      "reward": 0.4925,
      "reward_std": 0.21453479766845704,
      "rewards/compute_reward/mean": 0.4925,
      "rewards/compute_reward/std": 0.21453480839729308,
      "step": 300
    }
  ],
  "logging_steps": 50,
  "max_steps": 1000,
  "num_input_tokens_seen": 1331787,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}