kthierbach/unlimited-ocr-max

Card: state hardware by what is established, not by the dev host

Delete files model-00001-of-000001.safetensors with huggingface_hub

Upload README.md with huggingface_hub

Upload folder using huggingface_hub

card: --revision -> v0.2.1 (Hub-serving fix release; content otherwise unchanged)

card: int8 variant + decode throughput; --revision default to v0.2.0

card: document the int8 variant (model-int8.safetensors, GPU only)

weights: add int8 routed-expert variant (model-int8.safetensors)

model card: HF model-card guide pass

Unlimited-OCR MAX port: baidu weights unchanged, serve-ready config

initial commit