search

vasgaowei / pytorch_MELM

The pytorch implementation of the Min-Entropy Latent Model for Weakly Supervised Object Detection
104 stars 19 forks source link

How to get the loss in local min-entropy ? #2

Closed harrywang725 closed 5 years ago

harrywang725 commented 5 years ago

I saw your code in lib/net/network.py self._add_losses: https://github.com/vasgaowei/pytorch_MELM/blob/bd73dd11c938cb9256829ec3559daaab1fc77b74/lib/nets/network.py#L251-L254

But i saw the paper said the loss is that: (8(P$~J2B`F5Z~8(ED%T{{5

Is there something wrong?