Taehyeon Kim's picture

Taehyeon Kim

Kthyeon

·

https://kthyeon.com/

AI & ML interests

LLM Inference: Parallel, Speculative, Instructive Decoding

Recent Activity

authored a paper 8 days ago

AdaSTaR: Adaptive Data Sampling for Training Self-Taught Reasoners

authored a paper 8 days ago

$C^2$: Scalable Auto-Feedback for LLM-based Chart Generation

authored a paper 8 days ago

MERIT Feedback Elicits Better Bargaining in LLM Negotiators

View all activity

Organizations

New activity in microsoft/Phi-3-small-8k-instruct over 1 year ago

Getting the error: "triton.runtime.autotuner.OutOfResources: out of resource: shared memory, Required: 180224, Hardware limit: 166912. Reducing block sizes or `num_stages` may help."

#27 opened almost 2 years ago by

commented a paper almost 2 years ago

Towards Fast Multilingual LLM Inference: Speculative Decoding and Specialized Drafters

Paper • 2406.16758 • Published Jun 24, 2024 • 20 •