hao-ai-lab/Dynasor
[NeurIPS 2025] Simple extension on vLLM to help you speed up reasoning model without training.
This tool helps developers and ML engineers make their large language models (LLMs) respond faster and more efficiently, especially when those models are performing complex reasoning tasks. It takes an existing LLM setup, such as one running on vLLM, and optimizes its inference speed without needing any retraining. The output is a significantly faster and more resource-efficient LLM, which is particularly useful for applications requiring quick, thought-out responses.
224 stars. No commits in the last 6 months.
Use this if you are a machine learning engineer or developer looking to accelerate the performance of your LLM applications, especially those requiring complex reasoning, without investing time in model retraining or fine-tuning.
Not ideal if you need to fundamentally change an LLM's behavior or knowledge through training, rather than just optimizing its inference speed.
Stars
224
Forks
29
Language
Python
License
MIT
Category
Last pushed
May 31, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/hao-ai-lab/Dynasor"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
cvs-health/uqlm
UQLM: Uncertainty Quantification for Language Models, is a Python package for UQ-based LLM...
PRIME-RL/TTRL
[NeurIPS 2025] TTRL: Test-Time Reinforcement Learning
sapientinc/HRM
Hierarchical Reasoning Model Official Release
tigerchen52/query_level_uncertainty
query-level uncertainty in LLMs
reasoning-survey/Awesome-Reasoning-Foundation-Models
✨✨Latest Papers and Benchmarks in Reasoning with Foundation Models