Fetching metadata from the HF Docker repository...
-
Force CPU build with cmake args
-
Remove @spaces.GPU, CPU-only
-
Force CPU-only llama-cpp
-
Lazy-import llama_cpp inside @spaces.GPU
-
Add @spaces.GPU decorator for ZeroGPU
-
Add spaces package for ZeroGPU
-
Use pre-compiled llama-cpp wheel to avoid OOM
-
Remove Dockerfile
-
Revert to Gradio SDK
-
Add Dockerfile with pre-built binaries
-
Switch to Docker SDK
-
Add pre-built wheel for llama-cpp to avoid OOM
-
Add LLM chat app with LFM2.5
-
Add llama-cpp-python deps
-
initial commit
