pyine-v1-qwen3-4b-shortcut/reward_state.json

{
  "adapter": {
    "error_count": 0,
    "skip_count": 0,
    "total_count": 0
  },
  "manager": {
    "difficulty_estimator": {
      "bin_reward_stats": [
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        },
        {
          "count": 0,
          "max": NaN,
          "min": NaN,
          "sum": 0.0,
          "sumsq": 0.0
        }
      ],
      "bin_reward_values": [
        [],
        [],
        [],
        [],
        [],
        [],
        [],
        [],
        [],
        [],
        []
      ],
      "bin_term_reward_stats": {},
      "corr_stats": {
        "count": 0,
        "sum_x": 0.0,
        "sum_x2": 0.0,
        "sum_xy": 0.0,
        "sum_y": 0.0,
        "sum_y2": 0.0
      },
      "missing_count": 0,
      "override_skip_count": 0,
      "score_stats": {
        "count": 0,
        "max": NaN,
        "min": NaN,
        "sum": 0.0,
        "sumsq": 0.0
      },
      "score_values": [],
      "secondary_corr_stats": {},
      "secondary_stats": {},
      "total_count": 0
    },
    "epoch": 0.3998664886515354,
    "global_batch_counts": {
      "eval/": 1403,
      "train/": 600
    },
    "global_generation_counts": {
      "eval/": 280600,
      "train/": 576000
    },
    "local_generation_counts": {
      "eval/": 115,
      "train/": 0
    },
    "parsing_stats": {
      "answer_length_tokens": {
        "count": 0,
        "max": NaN,
        "min": NaN,
        "sum": 0.0,
        "sumsq": 0.0
      },
      "category_answer_length_tokens": {},
      "category_malformed_count": {},
      "category_missing_answer_count": {},
      "category_missing_reasoning_count": {},
      "category_output_length_tokens": {},
      "category_reasoning_length_tokens": {},
      "category_total_count": {},
      "malformed_count": 0,
      "missing_answer_count": 0,
      "missing_reasoning_count": 0,
      "output_length_tokens": {
        "count": 0,
        "max": NaN,
        "min": NaN,
        "sum": 0.0,
        "sumsq": 0.0
      },
      "reasoning_length_tokens": {
        "count": 0,
        "max": NaN,
        "min": NaN,
        "sum": 0.0,
        "sumsq": 0.0
      },
      "total_count": 0
    },
    "reward_category_stats": {},
    "reward_category_term_stats": {},
    "reward_term_stats": {
      "soft_match": {
        "count": 0,
        "max": NaN,
        "min": NaN,
        "sum": 0.0,
        "sumsq": 0.0
      }
    },
    "reward_total_stats": {
      "count": 0,
      "max": NaN,
      "min": NaN,
      "sum": 0.0,
      "sumsq": 0.0
    },
    "reward_total_values": [],
    "reward_total_values_count": 0,
    "step": null,
    "total_global_batch_count": 2003,
    "total_global_generation_count": 856600
  },
  "version": "0.1.4"
}