wllama.cpp multimodal demo (original) (raw)

Multimodal (Vision) Completion

Step 1: Load model

Downloads ~540MB from Hugging Face

Or pick local GGUF files (select both the LLM and mmproj file):

Step 2: Run completion

Prompt:

Image:

Generating…

Output:

Prompt: - t/s | Generation:- t/s