QeRL: Training 32B Models on Single H100 vs Three GPUs, Beating LoRA in Accuracy

16 October 2025
QeRL rainforcement learning quantization training speedup

QeRL: Training 32B Models on Single H100 vs Three GPUs, Beating LoRA in Accuracy

QeRL is a framework for training language models using reinforcement learning that simultaneously reduces GPU requirements and surpasses traditional LoRA and QLoRA methods in accuracy. On the Qwen2.5-7B-Instruct model, QeRL…

14 Free Courses on Machine Learning, Data Science, Data Analysis, and Python

30 August 2023
free machine learning data science analysis python courses

14 Free Courses on Machine Learning, Data Science, Data Analysis, and Python

The prevailing trend in online education is the rise of Massive Open Online Courses (MOOCs). Free courses on machine learning, data science, data analysis, and Python are readily available, based…

Robot manage objects from video tutorials using RL

19 February 2021

Robot manage objects from video tutorials using RL

In FAIR, the RL-agent was trained to manage objects using video tutorials. Standard RL algorithms are trained to a problem iteratively through learning from errors. The proposed algorithm learns a…

Evolving Simple Programs for Playing Atari Games

2 August 2018
Neural network playing Atari game

Evolving Simple Programs for Playing Atari Games

While a great number of researchers in Artificial Intelligence have focused their efforts on Deep Reinforcement Learning trying to beat human players on Atari games, researchers from the University of…