data/_run_metadata.json

{
  "input_dataset": "willwhim/wattsocr",
  "split": "train",
  "shuffle": true,
  "seed": 42,
  "max_samples": 20,
  "batch_size": 16,
  "source_dataset_sha": "e67bc770bfa459b86ed3760c43d996410f47fcc9",
  "temperature": 0.01,
  "top_p": 1e-05,
  "repetition_penalty": 1.1,
  "max_tokens": 8192,
  "task": "ocr",
  "gpu_memory_utilization": 0.8,
  "max_model_len": 8192,
  "total_batches": 2,
  "total_samples": 20,
  "model": "zai-org/GLM-OCR",
  "script": "glm-ocr-v2.py",
  "created_at": "2026-02-20T01:41:29.509129"
}