Fix firered-ocr.py: max_model_len 32768, handle existing column, JPEG encoding

Add FireRed-OCR script (2.1B, Qwen3-VL fine-tune)

Pin dots-ocr-1.5.py + ocr-vllm-judge.py to stable vLLM (>=0.15.1)

Fix dots-ocr.py OCR failures: pin vLLM, add limit_mm_per_prompt, per-image fallback

Add glm-ocr-v2.py: incremental uploads + checkpoint/resume

Restructure README: cleaner intro, TOC table, common options

Pin nanonets-ocr2.py + paddleocr-vl.py to stable vLLM (>=0.15.1)

Pin nanonets-ocr.py to stable vLLM (>=0.15.1) + datasets>=4.0.0

Pin rolm-ocr.py to stable vLLM (>=0.15.1) + datasets>=4.0.0

Update ocr-vllm-judge.py: improved eval prompt + --save-results flag

Switch hunyuan-ocr.py to stable vLLM (>=0.15.1), drop nightly index

Switch deepseek-ocr-vllm.py to stable vLLM (>=0.15.1), drop nightly index

Switch lighton-ocr2.py to stable vLLM (>=0.15.1), drop nightly index

Pin numarkdown-ocr.py to stable vLLM (>=0.15.1) + datasets>=4.0.0

Fix config_name=None breaking push_to_hub + use cu129 nightly index

Delete __pycache__

Add OCR vLLM judge script (jury mode, structured output)

Add --config/--create-pr support to dots-ocr.py

Add --config/--create-pr support to lighton-ocr2.py

Add --config/--create-pr support to deepseek-ocr-vllm.py

Add --config/--create-pr support to glm-ocr.py

Add --verbose flag to dump resolved package versions

Fix: use cu129 nightly variant (has x86_64 wheels)

Revert nightly index, add note: vLLM nightly x86_64 builds broken as of 2026-02-13

Drop vLLM nightly index, pin >=0.16.0rc (GLM-OCR in stable rc)

Trim prompt modes to official v2 ones (document, free)

Switch back to nightly vLLM, add addict + matplotlib deps

Use stable vLLM (>=0.15.1) instead of nightly for DeepSeek-OCR-2

Add DeepSeek-OCR-2 vLLM script (rewrite from fixed v1 pattern)

Pin datasets>=4.0.0, drop pyarrow pin

Pin pyarrow<18 to fix datasets compatibility

Fix deepseek-ocr-vllm.py: use official vLLM offline pattern (P0)

Update nanonets-ocr.py HF Jobs syntax and add smoke test notes

Replace fragile git dep with stable transformers>=5.1.0 for GLM-OCR

Fix glm-ocr.py max_tokens exceeding max_model_len

Remove extraneous f-string prefixes (ruff F541)

Replace deprecated HF_TRANSFER with HF_XET_HIGH_PERFORMANCE

Add hf-jobs tag to README frontmatter

Upload glm-ocr.py with huggingface_hub

Upload README.md with huggingface_hub

Upload glm-ocr.py with huggingface_hub

Upload glm-ocr.py with huggingface_hub

Upload glm-ocr.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub

Upload paddleocr-vl-1.5.py with huggingface_hub