AtomicChat/Unlimited-OCR-2-GGUF
Image-Text-to-Text • 3B • Updated
Unlimited-OCR-2
Baidu's 3.3B OCR model (page images to text with layout): GGUF for llama.cpp with v2 support, and W4A16 for transformers. By AtomicChat.
Note GGUF: BF16, Q8_0, AD-Q6/Q5/Q4 + mmproj. Needs llama.cpp with v2 support (see card).
AtomicChat/Unlimited-OCR-2-W4A16
Image-Text-to-Text • 3B • Updated • 1Note 4-bit weights (group 32) for transformers.