Convert PaddleOCR models to ONNX format
Evaluate model predictions with TruLens
Quantize a model for faster inference
Find and download models from Hugging Face
Teach, test, evaluate language models with MTEB Arena
Submit models for evaluation and view leaderboard
Display model benchmark results
Explore and visualize diverse models
SolidityBench Leaderboard
Calculate GPU requirements for running LLMs
Merge Lora adapters with a base model
Open Persian LLM Leaderboard
Convert Stable Diffusion checkpoint to Diffusers and open a PR
PaddleOCRModelConverter is a tool designed to convert PaddleOCR models into the ONNX format. ONNX (Open Neural Network Exchange) is an open standard that allows models to be transferred between different frameworks and platforms, enabling better interoperability and performance optimization. This tool is particularly useful for users who want to deploy PaddleOCR models in environments that support ONNX, such as.TensorRT, Core ML, or Edge Inference Engines.
• Model Conversion: Converts PaddleOCR models to ONNX format for cross-platform compatibility.
• Optimized Inference: Supports optimization of models for inference, ensuring faster and more efficient deployment.
• Framework Compatibility: Facilitates deployment across multiple ML frameworks and platforms.
• Command-Line Interface: Provides an easy-to-use command-line tool for model conversion.
• Cross-Platform Support: Enables deployment on diverse operating systems and hardware configurations.
paddle_ocr_model_converter --input_model path/to/model --output_path path/to/output
What models are supported by PaddleOCRModelConverter?
PaddleOCRModelConverter supports all standard PaddleOCR models, including but not limited to CRNN, Transformer, and PFAN models.
Why should I convert my PaddleOCR model to ONNX?
Converting to ONNX enables deployment in ONNX-compatible frameworks and platforms, which can improve inference performance and provide better interoperability.
Are there any specific dependencies required for the conversion?
Yes, ensure you have the latest versions of PaddlePaddle and ONNX runtime installed in your environment for smooth conversion and inference.