Philip-Bachman amdim-public issues

Philip-Bachman / amdim-public

Public repo for Augmented Multiscale Deep InfoMax representation learning

MIT License

396 stars 60 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Are amount of negative samples large?

#18 xinyaojiejie opened 1 year ago
0
Two pre-trained models are not downloading

#17 nanacoco419 opened 1 year ago
1
training self sueprvised learning on multi-gpu

#16 bhat-prashant opened 3 years ago
0
Fix the link to Kolesnikov et al.

#15 nzw0301 closed 4 years ago
1
time for training self-supervised and pretrained models

#14 georgecloudservices opened 4 years ago
1
Truly multi-view datasets

#13 bkj opened 4 years ago
1
Expectation of N7 in equation (1)

#12 martinmamql closed 2 years ago
2
urgent - how is positive and negative samples defined

#11 ghost opened 4 years ago
7
Why don't u use Resnet50 or any normal architectures?

#10 abcdvzz closed 4 years ago
1
Questions about Mixture-Based Representations

#9 Wuziyi616 closed 3 years ago
0
while in the test time, what dose the _warmup_batchnorm do?

#8 victor000000 opened 4 years ago
1
what does lgt_reg mean, is it the entropy maximization term H(q)?

#7 victor000000 closed 4 years ago
2
The effect of multi-scale features

#6 wangxu-scu opened 4 years ago
0
Details of Auto-augment to get the best results

#5 XingyuJinTI opened 4 years ago
4
Question about training self-supervised model

#4 0three closed 4 years ago
5
Which python command are you using for CIFAR and STL10 experiments?

#3 LIUNancy closed 4 years ago
6
How to avoid the grad from labels influencing the info_modules?

#2 munanning closed 5 years ago
2
Clean checkpoints

#1 wbuchwalter closed 5 years ago
1