evit

evit copied to clipboard

Awesome works! I want to know how Figure 5 is generated.Is it by putting the mask back on the original image ?

wyl753

Awesome works! I want to ask that whether you have experiment your method under more extreme MACs drop ratio, such as the accuracy drop under pruning 80% of MACs. Thanks!

King4819

guspan-tanadi

I would like to ask, if I use oracle vit to train evit, does it still need to train 300 epochs?

Summer-nan

How long does it take to retrain in single A100 sheets if no pre-trained model is used? 150 hours? How do I understand the reference in the article to using...

Summer-nan

Can you share another model trained by evit? (LV ViT, DeiT-T, B, etc..) Thanks

jhtwosun

cls_attn = attn[:, :, 0, 1:]

4

hi，thank you for your contributions to open source. but i have a problem. cls_attn = attn[:, :, 0, 1:] What does this code mean？

Bugs-Bunny01

Visualize

1

Hello, I am using this command, but it does not save any masked images. python3 main.py --model deit_small_patch16_shrink_base --fuse_token --base_keep_rate 0.7 --visualize_mask --n_visualization 64 --resume checkpoint --data-path /path/to/imagenet I think...

bsaberla

When to start fusing image patches

1

hi, I now have a question, do you start to fuse image blocks from the first epoch, will there be a large error

Bugs-Bunny01

tokens是什么，为什么在evit.py里赋值none

1

File "C:\Users\17664\Desktop\timesformer - evit2\src\grasping\Models\timeSformer_orig_four.py", line 600, in forward_features_tactile_sliding x, left_token, idx = blk(x, keep_rate[i], tokens[i], get_idx) TypeError: 'NoneType' object is not subscriptable 首先感谢您为视觉开源社区做出的贡献，不过我将您的融合图像块模块添加到我的模型里时报这个错误，tokens被赋值none，为什么显示'NoneType' object is not subscriptable呢？ 我的最大疑惑还是tokens到底是什么，为什么在evit.py文件里一直赋值none。tokens是所有的图像块吗？

Bugs-Bunny01