Hugging Face

Hugging Face adds native GGUF inference to Transformers, starting with Qwen3.5 on Macs

Transformers now loads GGUF models in packed form and runs them on Apple Silicon GPUs via Metal kernels, starting with Qwen3.5.