harleyszhang/rapid_llm

RapidLLM is an LLM inference framework with continuous batching, tensor and data parallelism, quantized weights, and with pluggable Triton/CUDA kernels.

Python195 stars33 forks

attention continues continuous-batching flash-attention-3 llama3 llm llm-inference python3 qwen3-moe qwen3-vl tensor-parallel triton-kernels

Momentum

harleyszhang/rapid_llm has 195 GitHub stars.

It currently ranks #32711 daily, #33047 weekly, #34480 monthly by momentum on Trending Repos; its best-ever rank is #13816.

Trending Repos first recorded it on Sep 30, 2026.

Rankings

PeriodRankScore

Daily

#327110.0

Weekly

#330470.0

Monthly

#344800.0

How scoring works

Engagement

Total stars

Daily new stars

Embeds

Best rank #13816

Format

Markdown

[![harleyszhang/rapid_llm badge](https://trending-repos.com/badge/harleyszhang/rapid_llm.svg?v=3)](https://trending-repos.com/repositories/harleyszhang/rapid_llm?utm_source=github&utm_medium=badge)

harleyszhang/rapid_llm

Momentum

Rankings

Engagement

Similar trending repositories

Embeds

harleyszhang/rapid_llm

Momentum

Rankings

Engagement

Similar trending repositories

Embeds