home/categories/machine-learning/orchestra-research-ai-research-skills-06-post-training-verl-skill-md
machine-learningdata-ai
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
maintainer
Orchestra-Research
Обновлено 1/29/2026
Звёзды
6563
Форки
515
quick start
Installation and usage
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
Установка
$ install --globalskills.sh
Использование
После установки вы можете использовать этот skill, выполнив следующую команду в терминале:
skills use verl-rl-training