Issues
- 1
A problem about the PSA processing process
#47 opened by 0524hhh - 6
MCTformer+ post-processing parameter problem
#42 opened by luzedong - 7
Reproduction of results
#37 opened by pipizhum - 2
MS COCO 2014 dataset
#41 opened by jungletada - 1
Code to the paper "Learning Multi-Modal Class-Specific Tokens for Weakly Supervised Dense Object Localization" not available
#44 opened by Billy-06 - 3
RuntimeError
#43 opened by geek-APTX4869 - 8
- 0
- 0
Yes, good question.
#48 opened by whb919168848 - 0
SBD dataset seems different to the origin one.
#46 opened by Matrix53 - 0
How to convert CAM to segmentation labels?
#45 opened by BuyiZhu - 1
code of Learning Multi-Modal Class-Specific Tokens for Weakly Supervised Dense Object Localization.
#39 opened by Nastu-Ho - 1
how to make our own cls_label.npy
#40 opened by mengweibao - 1
- 1
run.sh problem
#36 opened by pipizhum - 1
How do you validate the results of your seed cam on the val set of VOC, and can you provide your program?
#38 opened by massica17 - 21
Reproducing the results of segmentation
#26 opened by Yingyue-L - 6
How to train COCO?
#15 opened by zhaoxiaodong789 - 5
Does COCO dataset need psa and seg step?
#25 opened by vtars - 2
- 1
How to generate 'train_aug_id.txt' for COCO?
#35 opened by tg-Flipped - 1
PSA+DeeplabV1 or IRN+DeeplabV2?
#27 opened by guangpintao - 8
Performance on DEIT-base model
#31 opened by Ferenas - 0
train_aug_id.txt
#30 opened by 111chengxuyuan - 0
Training own dataset about res38_cls
#34 opened by dt9640214 - 0
multi-class tokens on CUB or ImageNet
#32 opened by Nastu-Ho - 2
about run.sh
#28 opened by pppwzj - 0
ERROR RuntimeError: shape '[-1, 13, 13, 384]' is invalid for input of size 67968
#29 opened by pppwzj - 1
Performance Concern.
#23 opened by guangpintao - 2
Training our own dataset
#24 opened by xiaona123456 - 1
Evaluation of the generated class-specific activation map results does not reach the level of the ablation experiments in the paper
#22 opened by yijiangWang0316 - 4
I have a question about engine.py
#21 opened by yijiangWang0316 - 1
I have a question about engine.py
#20 opened by Nastu-Ho - 1
The results of seed cam can't achieve 61 mIoU
#18 opened by sennnnn - 4
About the optimizer details
#2 opened by PlumedSerpent - 1
the multi-token can't train the good result
#4 opened by YuYue26 - 1
Informative: error triggered by package gimm
#19 opened by fish98 - 2
run_seg.sh command line parameters
#11 opened by AmeenAli - 1
- 1
The hyper-parameters in PSA step?
#9 opened by PlumedSerpent - 0
Looking forward to ur codes~
#6 opened by Roc-nn - 0
- 0
Waiting for your code!
#3 opened by yangyang99 - 1
Looking for your implement~
#1 opened by Unrealluver - 1
I am wondering if anyone sucessfully reproduce the results claimed in the paper now?
#16 opened by zbf1991 - 2
some problem about loading pretrained model
#12 opened by zaiquanyang - 3
What is "init_weights" in command parameters to train the segmentation model?
#13 opened by guangpintao - 0
- 2
no module named 'utils'
#7 opened by YuYue26