pytorch / pytorch/rl

[Feature Request] DQN Loss importance sampling

Open
#955 2 comments 0 reactions 1 assignee View on GitHub

@vmoens is already working on this.

Since Mar 8, 2023.

enhancement
Dominant language
Python
Stars
3.6k
Forks
484
Avg merge
1d 1h
Merged PRs (30d)
207

Description

Motivation

I believe the current DQN Losses don't apply importance sampling weights. This is almost always applied when using a PER Buffer.

Solution

The PER buffer outputs "_weight" in info, this if in inputTensorDict, could be applied at time of loss calculation

Alternatives

element-wise loss could be returned with any transforms or weighting to be applied later on

Checklist

  • I have checked that there is no similar issue in the repo (required)

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.