deepseek-ocr

Skill by ara.so — Daily 2026 Skills collection. DeepSeek-OCR is a vision-language model for Optical Character Recognition with "Contexts Optical Compression." It supports native and dynamic resolutions, multiple prompt modes (document-to-markdown, free OCR, figure parsing, grounding), and can be run via vLLM (high-throughput) or HuggingFace Transformers. It processes images and PDFs, outputting structured text or markdown. Model is available on HuggingFace: deepseek-ai/DeepSeek-OCR Edit DeepSeek-OCR-master/DeepSeek-OCR-vllm/config.py : vLLM 0.8.5 requires transformers>=4.51.1 — if running both in the same env, this error is safe to ignore per the project docs. Ensure NGramPerReqLogitsProcessor is passed to LLM — this is required for proper decoding: Add table token IDs to the whitelist: - CUDA 11.8+, PyTorch 2.6.0 - Python 3.12.9 (via conda recommended) - Use smaller resolution: base_size=512 or base_size=640 - Disable crop_mode=False to avoid multi-crop dynamic resolution - Reduce batch size in vLLM inputs