local llm on laptop 780M GPU using llama + gemma 4 qat

06 Jun, 2026

1. install llama.cpp https://llama.app/
2. llama serve -hf unsloth/gemma-4-12B-it-qat-GGUF
   this gives me error: Failed to load CLIP model from
   llama serve -hf unsloth/gemma-4-12B-it-qat-GGUF --no-mmproj
   https://unsloth.ai/docs/models/gemma-4/qat
3. Open http://127.0.0.1:8080 test if it works Performance: 9 tokens/sec using Vulkan backend Next steps: try to use opencode with local llm

#llm