Flint: Efficiently Leveraging High Bandwidth Flash for LLM Inference

(arxiv.org)

2 points | by metrofun 4 hours ago ago

No comments yet.