SimpleMLP Dropout probability

[zafstojano]

Dear Avalanche community,

I believe I have found a tiny bug in the SimpleMLP class.

In particular, the drop_rate from the __init__ method is not passed to nn.Dropout for the hidden layers of the MLP.
In turn, this instantiates a Dropout layer with default probability p=0.5.

I have created a simple pull request solving this issue.

Please let me know if I should perform anything differently.

Best regards,
Zafir

[vlomonaco]

Thanks @zaf-stojano! It was merged in #896! :)

[zafstojano]

Cheers 🍻