Comments (1)
Hi,
Thanks for your question! Regarding the BN momentum, we followed the hyperparameters in the previous work like EMANet. This is an old-school implementation, mainly because we finetune the ImageNet-pretrained backbone for the segmentation task with limited training samples and small batch size.
According to my benchmark, it makes no difference for HamNet to adopt the classic 0.1 BN momentum under the ablation settings but shifts the peak validation performance to an earlier time. Maybe it can slightly improve the average performance or shorten the training schedule with the 0.1 BN momentum. This 0.1 BN momentum has already been applied to my collaborator Meng-hao Guo's new paper, External Attention.
Please let me know if you have additional questions. Thank you!
from enjoy-hamburger.
Related Issues (10)
- Applying Hamburger to other models makes the training collapse soon HOT 2
- Confusion on update rules of Ham HOT 1
- About Fix point iteration HOT 3
- question about the implementation of one-step gradient
- Arxiv version or blog HOT 5
- KeyError: 'van_tiny is not in the models registry' HOT 4
- 缺少主干网络 HOT 2
- Default process group has not been initialized HOT 4
- Difference between the code and Eq(13) in the paper about the gradient calculation HOT 5
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from enjoy-hamburger.