Thomas Rochefort-Beaudoin

[Bug] GPU memory explodes when using Conv2D layers in Dict Observations FeatureExtractor

9

### 🐛 Bug I am having issues in SB3 with a CustomFeatureExtractor for a Dict observation space that is making my GPU memory explode. The observation space is composed of...

- Parser type:Article - http://quote.morningstar.ca/Quicktakes/Financials/is.aspx?t=AAPL&region=USA&culture=en-CA&ops=clear, - http://quote.morningstar.ca/Quicktakes/Financials/is.aspx?t=ABX&region=CAN&culture=en-CA&ops=clear, - http://quote.morningstar.ca/Quicktakes/Financials/is.aspx?t=AMZN&region=USA&culture=en-CA&ops=clear

- Parser type:Table - http://quote.morningstar.ca/Quicktakes/Financials/is.aspx?t=AAPL&region=USA&culture=en-CA&ops=clear, - http://quote.morningstar.ca/Quicktakes/Financials/is.aspx?t=ABX&region=CAN&culture=en-CA&ops=clear, - http://quote.morningstar.ca/Quicktakes/Financials/is.aspx?t=AMZN&region=USA&culture=en-CA&ops=clear

Number of ResNet block

3

@OrigamiDream The paper mentions on page 4 "Tokens belonging to image patches for any time-step are embedded using a **single** ResNet block". So we could update the Readme.md and just...

Dataset help

3

Hey @OrigamiDream ! I am also looking into implementing the Gato paper. In order to complement your work, I decided to start looking into the datasets mentioned in the paper...