search

uqzhichen / GSMFlow

[ACM MM 2021] Mitigating Generation Shifts for Generalized Zero-Shot Learning
6 stars 3 forks source link

Missing self. ? #3

Closed miguelvalente closed 2 years ago

miguelvalente commented 2 years ago

Is there a self. missing? Otherwise the _trainatt variable is never used

https://github.com/uqzhichen/GSMFlow/blob/5db88aa806b944a9753758d48a59dbf2b2eba7d2/dataset_GBU.py#L32

miguelvalente commented 2 years ago

Same thing with this variable. Its never used

https://github.com/uqzhichen/GSMFlow/blob/5db88aa806b944a9753758d48a59dbf2b2eba7d2/dataset_GBU.py#L60

uqzhichen commented 2 years ago

Hi there,

As you can see, the variable seen_pro is used instead here as the seen class attributes. Same for training labels, train_idx is used instead as well.

This repo is not for developing an off-the-shelf system, if there are more typos in it, I feel sorry about the impect souce code.

Cheers, Zhi

miguelvalente commented 2 years ago

Thanks! I understand the repo is not dev oriented. The way things are set up makes the code and mat files matching confusing. I've lost more than a couple of hours making sense of it. If anything these issues might help someone who is going through the same process!

Nonetheless, thx for making the code for the paper public. BR, Miguel