Download and test one Hugging Face LLM at a time. Auto routes
standard Transformers checkpoints to Transformers and GGUF files to
the CUDA-enabled llama.cpp backend. GGUF repositories are inspected
first so you can select only the Q4/Q5/Q8 (or newer) quant you want.
GGUF quant file (choose one)
Disk cache: no model selected.
Status: inspect a repository, then Download and Load it.