pytorch/examples

Actor critic example not using discount rate properly

Open

#744 建立於 2020年3月27日

在 GitHub 查看
 (3 留言) (0 反應) (0 負責人)Python (21,634 star) (9,429 fork)batch import
good first issuetriaged

描述

The Actor Critic example (which is actually an implementation of REINFORCE-with-baseline as pointed out in https://github.com/pytorch/examples/issues/573), does not use the discount rate properly.

The loss should include \gamma ^ t, as shown in the box on page 330 of Sutton & Barto:

image

貢獻者指南

技術棧
pythonpytorch
領域
machine learningai
議題類型
bug
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
REINFORCE algorithmPyTorch basicsDiscount factor concept
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
45
研究方向
Examine the actor critic example file (likely in the reinforcement learning folder). The current code likely computes the loss without weighting by gamma^t. Refer to the referenced issue #573 and the Sutton & Barto equation. The fix involves multiplying each log probability term by gamma^t in the policy gradient loss calculation. Check the comments in the issue for any additional guidance.