Modelpublic

fernandotonon/QtMeshEditor-smolvlm-gguf

sourceHugging Faceapache-2.0updated 3mo agoView on Hugging Face

1likes60downloads

Model Card

SmolVLM-500M-Instruct — Q8_0 GGUF

Quantized [HuggingFaceTB/SmolVLM-500M-Instruct](https://huggingface.co/HuggingFaceTB/SmolVLM-500M-Instruct) (Apache-2.0) as a Q8_0 GGUF + vision projector for llama.cpp inference. All credit for the model goes to the Hugging Face TB team.

Packaged for [QtMeshEditor](https://github.com/fernandotonon/QtMeshEditor)'s local image-captioning feature (llama.cpp runs in-process; the model downloads on first use and runs offline).

The files QtMeshEditor downloads at runtime live in the shared `fernandotonon/QtMeshEditor-models` repo under caption/. This repo is the standalone model card + mirror.

Files

file	role
`SmolVLM-500M-Instruct-Q8_0.gguf`	language model, Q8_0 (~437 MB)
`mmproj-SmolVLM-500M-Instruct-Q8_0.gguf`	vision projector (~109 MB)

Use with llama.cpp's multimodal API (load the mmproj alongside the LM, e.g. llama-mtmd-cli -m SmolVLM-500M-Instruct-Q8_0.gguf --mmproj mmproj-SmolVLM-500M-Instruct-Q8_0.gguf).

License

Apache-2.0 (same as upstream). Credit: Hugging Face TB (SmolVLM).