shreybirmiwal

Joined 4 years ago

Karma

21

Submissions

11

Member since

Oct 29

View on Hacker News

Recent submissions

SparklingTree: 30-40% faster specdec than DSpark by combining DDTree and DSpark jwlabs.vercel.app

2 shreybirmiwal 3 months ago 0 comments

Comment · 3 months ago

LLM speculative decoding explained simply jwlabs.vercel.app

2 shreybirmiwal 3 months ago 0 comments

Comment · 3 months ago

thank you so much

Predictive Speculative KV Replication for Bursty LLM Inference jwlabs.vercel.app

41 shreybirmiwal 3 months ago 4 comments

Specializing Speculative Decoding to increase drafter accuracy jwlabs.vercel.app

2 shreybirmiwal 3 months ago 0 comments