Train, fine-tune, and export PaddleOCR models for text detection and recognition.
npx skills add https://github.com/jayll1303/AIEKit --skill paddleocr-jayll1303
Please help me install this Agent Skill. Skill: paddleocr Source: https://github.com/jayll1303/AIEKit/tree/main/.kiro/skills/paddleocr Command: npx skills add https://github.com/jayll1303/AIEKit --skill paddleocr-jayll1303
Provides a complete, actionable workflow to train, fine-tune, export, and run high-quality OCR models with PaddlePaddle's PaddleOCR so teams can convert images and documents into accurate structured text without ad-hoc tooling gaps.
Ask the skill to install PaddlePaddle and PaddleOCR, prepare a PaddleOCR-formatted train set, fine-tune a detection or recognition model on that data, export the inference model, and run a sample inference on one image.
High-intent search queries and answers about installing and using this skill.
Fine-tune PaddleOCR by preparing a labeled dataset, selecting pretrained detection or recognition models, modifying YAML training configs, and applying learning-rate scaling rules for multi-GPU domain adaptation.
Prepare PaddleOCR datasets by formatting detection and recognition labels, using PPOCRLabel for annotation, and managing dictionaries to support multilingual training requirements.
Export PaddleOCR to ONNX by first converting the trained model to inference format, then enabling high-performance inference settings, and finally performing ONNX conversion for deployment.
PaddleOCR supports document parsing for receipts, IDs, and forms by utilizing PP-StructureV3 to detect and extract structured text from multilingual document images.
Fix PaddleOCR training OOM errors by applying Automatic Mixed Precision (AMP) tips, reducing batch sizes, and following memory handling best practices provided in the troubleshooting guidance.
Integrate PaddleOCR inference with MCP or Triton serving by exporting the model to inference format, enabling high-performance inference, and configuring the serving integration for production workflows.
dongzhuoyaoConfigure Slurm and ROCm environments for PyTorch jobs on LUMI.
doanchienthangdevDeploy machine learning models with TensorFlow Serving, TorchServe, or Triton Inference Server.
munluckyExtract and verify Korean and English text from images using OCR engines.
TopduExtract text, formulas, and tables from images and scanned PDFs.
