RL algorithm: from PPO to GRPO and DAPO

RL algorithm: from PPO to GRPO and DAPO

https://zhuanlan.zhihu.com/p/1898817630208517687

$\text{Latex Example}\quad {\color{green}green}\quad {\color[rgb]{0.286,0.529,0.808} light-blue}\quad {\color[rgb]{0.553,0.133,0.537}purple} {\color[rgb]{0.820,0.208,0.208}\quad light- red}\quad {\color{brown}brown}$ Read more ›

More Articles

• conda, pip, docker: source
• The State of Reinforcement Learning for LLM Reasoning (1)
• A (Long) Peek into Reinforcement Learning: Part3
• conda, pip, docker: source
• A (Long) Peek into Reinforcement Learning: Part2
• A (Long) Peek into Reinforcement Learning: Part1
• Spinning Up: Part 3: Intro to Policy Optimization
• Spinning Up: Part 2: Kinds of RL Algorithms
• Spinning Up: Part 1 Key Concepts inb RL
• Docker, sftp, conda
• Post with Image and Caption
• Paper List: RL-Strawberry
• The Surprising Effectiveness of Test-Time Training for Abstract Reasoning
• NN-Z2H: Lecture 3: Building makemore Part 2: MLP
• NN-Z2H: Lecture 2: The spelled-out intro to language modeling: building makemore
• Show All Articles ›