Last released Jul 18, 2025
A linear-attention transformer implementation with KV caching.
Supported by