eval_set_summary.json

{
  "name": "ocr_document_processing_eval",
  "description": "Document-processing OCR proxy for digitization, KYC-like numeric fields, and RAG extraction checks.",
  "task": "document_processing_ocr",
  "output_dir": "data/eval_ocr/ocr_document_processing_eval",
  "ocr_rows": 648,
  "sharegpt_rows": 648,
  "image_errors": 0,
  "images_per_dir": 5000,
  "outputs": {
    "ocr_jsonl": "data/eval_ocr/ocr_document_processing_eval/ocr_document_processing_eval.ocr.jsonl",
    "sharegpt_json": "data/eval_ocr/ocr_document_processing_eval/ocr_document_processing_eval.sharegpt.json",
    "dataset_info": "data/eval_ocr/ocr_document_processing_eval/dataset_info_fragment.json",
    "metadata_jsonl": "data/eval_ocr/ocr_document_processing_eval/ocr_document_processing_eval.metadata.jsonl"
  }
}