languages/chinese/config.json

{
  "model_name": "PP-OCRv5_server_rec",
  "model_type": "recognition",
  "framework": "PaddleOCR",
  "version": "PP-OCRv5",
  "language_group": "chinese",
  "supported_languages": [
    "Chinese (Simplified)",
    "Pinyin",
    "Traditional Chinese",
    "English",
    "Japanese"
  ],
  "description": "PP-OCRv5 Chinese/Japanese text recognition model (server version)",
  "accuracy": "N/A",
  "dataset": "Default PP-OCRv5 multilingual dataset",
  "original_format": "PaddlePaddle",
  "converted_format": "ONNX",
  "opset_version": 11,
  "dictionary_file": "ppocrv5_dict.txt",
  "input_shape": "dynamic (batch_size, 3, 32, dynamic_width)",
  "output_shape": "dynamic (batch_size, sequence_length, num_classes)",
  "source": "https://github.com/PaddlePaddle/PaddleOCR",
  "documentation": "https://www.paddleocr.ai/main/en/version3.x/algorithm/PP-OCRv5/PP-OCRv5_multi_languages.html",
  "usage": {
    "library": "onnxruntime",
    "preprocessing": "Grayscale or RGB image, resized to height=32, normalized",
    "postprocessing": "CTC decoding with character dictionary",
    "dictionary_format": "One character per line, UTF-8 encoding"
  }
}