As of today, 2026/03/20, not even a simple and lightweight model like bartowski/Qwen_Qwen3.5-0.8B-GGUF works, giving "Error loading Qwen_Qwen3.5-0.8B-Q4 0.gguf: Unsupported model architecture: qwen35".
Like countless other models downloaded from Hugging Face.
Why allow this feature if not even a single one works?
As of today, 2026/03/20, not even a simple and lightweight model like bartowski/Qwen_Qwen3.5-0.8B-GGUF works, giving "Error loading Qwen_Qwen3.5-0.8B-Q4 0.gguf: Unsupported model architecture: qwen35".
Like countless other models downloaded from Hugging Face.
Why allow this feature if not even a single one works?