Lightlines
By reason-machines

turboquant-pytorch

reason-machines814 installs

Smaller LLM memory caches with accurate attention.

Recent revisions are kept. Restoring one creates a new revision.