Topic

Hybrid Attention

1 items across the graph — tagged with Hybrid Attention.

From the graph · 1

repo
marzukia/qMLX

qMLX: Custom inference engine for Qwen 3.5 122B on Apple Silicon, extending MLX with hybrid attention support, SSD-backed KV cache, and RYS layer duplication fo…

→

Related topics

inference 1hybrid-attention 1llm 1mlx 1mtp 1kv-cache 1rys 1qwen 1apple-silicon 1
Search Hybrid Attention →All topics →