happynew111/haotian_data-GPS-verl-main-CL

AdaRFT: Adaptive Curriculum Reinforcement Finetuning 📢 New extension to verl! We propose an adaptive curriculum learning method for efficient and scalable reinforcement finetuning (RFT) of LLMs — now implemented as an extension to this repo. Efficient Reinforcement Finetuning via Adaptive Curriculum LearningTaiwei Shi†, Yiyang Wu†, Linxin Song†, Tianyi Zhou▽, Jieyu Zhao††University of Southern California, ▽University of Maryland[Paper] 🧠 Highlights Dynamically adapts… See the full description on the dataset page: https://huggingface.co/datasets/happynew111/haotian_data-GPS-verl-main-CL.