Upload qwen3.5-0.8b.mq6 with huggingface_hub

update README: add MQ4 entry, refresh perf table with 7900 XTX numbers + new naming

add qwen3.5-0.8b.mq4 (MagnumQuant 4-bit, FWHT-rotated HFQ4-G256)

Rename to HF4/HF6 naming convention (hipfire v0.1.3)

Upload README.md with huggingface_hub

revert to original quantizer output filenames

rename: {family}-{size}.{quant} convention, drop experimental variants

Upload qwen3.5-0.8b.hfq6.hfq with huggingface_hub

Upload qwen3.5-0.8b.q4.hfq with huggingface_hub

initial commit