get a different loss and accuracy curve

[qimw]

Hi, and I am using the default parameters training the net, but I got some different loss and training accuracy curves compared with the README file. Has anyone meet this kind of things?

• smallNORB all loss:
• smallNORB training accuracy:
• mnist all loss:
• mnist training accuracy:

[yhyu13]

@qimw @www0wwwjs1

Your results looks pretty good.

I apologize for the confusion, the learning curves we proposed include weight decay actually. However, by default, the weight decay loss is set to false (here). From your result, we can tell that weight decay is a essential regularization.

[qimw]

Hi, thank you for your prompt reply : )