Modelpublic

Mintzs/ora-qwen3.5-4b-docsynth-v1-GGUF

sourceHugging Faceotherupdated 11d agoView on Hugging Face

0likes143downloads

Model Card

Ora DocSynth v1 GGUF

GGUF exports of Mintzs/ora-qwen3.5-4b-docsynth-v1 at revision 3d719e0f6b0e8364624e905d4068c77e8148e52d.

Files

—Ora-DocSynth-v1-Q4_K_M.gguf: smaller local conversational model.
—Ora-DocSynth-v1-Q8_0.gguf: higher-fidelity, larger local model.
—mmproj-Ora-DocSynth-v1-F16.gguf: optional vision projector for compatible llama.cpp multimodal runtimes.

The text trunk was converted with --no-mtp and both quantizations were derived directly from the same F16 GGUF. Converter revision: 6a1a922d269908a29cbd4b49c27e6a8e7fd10fae. See conversion_manifest.json for checksums and validation status.

Scope and restriction

Experimental research-only, non-commercial release. The training dataset includes data derived from NVIDIA ChatQA and is not cleared here for commercial client deployment.