self-hosted/ai
§01·model · /models

llava 7b

multimodalactiveLlama 2 Community License

7B open vision-language model (LLaVA 1.5) by Liu et al. - a Vicuna/Llama-2 LLM paired with a CLIP visual encoder for image understanding and VQA. Llama 2 Community License.

§02·GPUs that run this model
1 total
GPUVRAMSeriesBest speedMin VRAMWorksBenchmarksRecipe
RTX 3060 Ti8GB3072tokens/s8GB1recipecheck ↗

benchmarked·~ runs via recipe (not benchmarked)· untested·doesn't fit