OvisOCR2 BF16 for PDF Bookmark Forge

Unquantized BF16 GGUF conversion of ATH-MaaS/OvisOCR2 revision

65c619d374b55d4152e85150fc1b003700bc1f0c, produced and tested with

llama.cpp b10373 commit 38406d597fe651aafc7d843d270cb25b5ec2617a.

Files:

OvisOCR2-BF16.gguf — SHA-256 1ca4a12cfad12a7d5474a23de43366c6d9efc98249ffa69d76a326f3543b21b2
mmproj-OvisOCR2-BF16.gguf — SHA-256 2d3eba7be09d7d942ab6f82f126f32aa268be74bd6f9d05e9b176addc1f4126b

The main conversion uses --no-mtp because the upstream configuration declares

one speculative MTP layer whose tensors are not present in the published weights.

This does not quantize the model or remove OCR model layers.

Upstream model and this converted distribution are Apache-2.0. See the upstream

model card for intended use and limitations.

Question 2

What license applies to MDfox/OvisOCR2-BF16-GGUF-b10373?

Accepted Answer

License: apache-2.0. Verify terms on Hugging Face before commercial use.

Question 3

How do I run MDfox/OvisOCR2-BF16-GGUF-b10373 locally?

Accepted Answer

Download a GGUF file from this page and load it in guIDE or llama.cpp. Pipeline task: text-generation.

Question 4

How much VRAM or disk space does MDfox/OvisOCR2-BF16-GGUF-b10373 need?

Accepted Answer

Runs locally from ~197.7 MB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).

MDfox/OvisOCR2-BF16-GGUF-b10373 overview

File	Type	Quantization	Size	Link
OvisOCR2-BF16.gguf	GGUF	BF16	1.41 GB	Download
mmproj-OvisOCR2-BF16.gguf	GGUF	BF16	197.7 MB	Download