sense-x / tokenmix Goto Github PK
View Code? Open in Web Editor NEWTokenMix: Rethinking Image Mixing for Data Augmentation in Vision Transformers (ECCV 2022)
License: MIT License
TokenMix: Rethinking Image Mixing for Data Augmentation in Vision Transformers (ECCV 2022)
License: MIT License
Hi, I am curious about the problem of dimension inconsistency.
(1) The shape of "score_map" that generated in generate_label.py is [2, 5, H, W], but the dimension of score_maps seems to be [2, H, W, 5] in "score_maps[-1,0,0,5]=target " (line 68 of TokenMix)/relabel/dataset.py )
(2) The dimension of "label_maps_topk" in line 48 of TokenMix/tokenmix.py is [batch_size, 3, H, H, 5], but I cannot find how to transform "score_maps" to "label_maps_topk", and what information is stored in the 0, 1, 2 dimension of "label_maps_topk", respectively.
The paper briefly mentioned that uniform random TokenMix is ineffective for CNNs but beneficial for Transformers. However, I did not find empirical numbers for these claims. Could you provide the numbers for comparison with CutMix as well as structured random TokenMix?
In addition to TokenMix before the first Transformer block, have you considered or tried TokenMix in the middle of the model?
hi i'm very interested in the data augmentation which you release.
can you release the code which is about tokenmix add in to swin transformer?
thank you alot
Dear @jihaonew
How we can run TokenMix on CIFAR10 dataset?
Hello guys,
I am Bill from ECCV. We had a nice talk about your work during your poster session.
Congrats for one more time :)
I would like to ask whether you could provide us the checkpoints of your pretrained models.
Thanks a lot ;)
When is the source code going to be released?!
Very sorry, probably because my code ability is too weak and did not find the code about the neural activation map part, can you tell me how it manifests.
Thank you very much
What is the basis or prior knowledge of your partition?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.