Cerenaut / pt-aha

AHA implemented in PyTorch. New features are being implemented here (not in the TF version).
Apache License 2.0
0 stars 2 forks source link

Modifications necessary to make AHA work with cls_module configured w… #1

Closed affogato closed 3 years ago

affogato commented 3 years ago

…ith the VGG implementation that was brought over from CFSL.

affogato commented 3 years ago

No, I intended to remove the commenting. I added the check elsewhere.

On Tue, 23 Feb 2021 at 15:33, Abdelrahman Ahmed notifications@github.com wrote:

@abdel commented on this pull request.

In cls_module/cls_module/memory/ltm/vgg.py https://github.com/Cerenaut/pt-aha/pull/1#discussion_r580762764:

@@ -87,6 +81,8 @@ def forward_memory(self, inputs, targets, labels):

 preds, encoding = self.vgg.forward(x=inputs, num_step=0, training=self.training, return_features=True)
  • if not a list, then don't need idx, if list, use index

was there meant to be check here?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Cerenaut/pt-aha/pull/1#pullrequestreview-595989827, or unsubscribe https://github.com/notifications/unsubscribe-auth/AADPIHSLDXZCHWAFTSFMM7LTAMVYLANCNFSM4YBXROJQ .

-- Gideon Kowadlo http://gideon.kowadlo.net