QVAC Fabric LLM in the next weeks and months will continue to evolve to support local on-device inference at scale across any GPU, OS and device. From our latest paper https://huggingface.co/blog/qvac/fabric-llm-finetune - On-Device Fine-Tuning on Mobile GPUs: We demonstrate the first successful fine-tuning on mobile GPUs (Adreno, Mali, Apple), a previously unsupported capability that unlocks true on-device personalization. - Seamless Cross-Platform Scalability: Our solution provides universal compatibility across the entire desktop GPU ecosystem, including AMD, Intel, NVIDIA, and Apple architectures.(Paolo Ardoino 🤖)