Unbiased GRPO Optimization: Dr. GRPO

Medium · Solved in PyTorch · Machine learning coding practice

Problem

Implement the Unbiased Group Relative Policy Optimization (Dr. GRPO) objective function.

Topics: RL, policy-optimization

The full statement, worked examples, hints and the test suite are available once you sign in. You can then solve Unbiased GRPO Optimization: Dr. GRPO in the browser and run it against the tests.

Related problems

• DAPO: Clip-Higher Policy Optimization Loss Medium
• Temporal Difference (TD) Error for Action-Value Function Easy
• Optimizers: From Momentum to Adam Medium
• AdamW Optimizer Medium
• Implement Bagging Classifier with Decision Stumps Medium
• Basic KV Cache: Prefill and Decode Medium