Comments (3)
The MIC repository does not support source-only training at the moment.
from mic.
The MIC repository does not support source-only training at the moment.
Training on a source-only dataset might not get a good result, right? I think MIC method is suitable for domain adaptive problems, am I right? Looking forward to your reply!
from mic.
Training only on the source domain will naturally perform worse than utilizing UDA with additional unlabeled target images. You can find some reference results in the DAFormer paper https://arxiv.org/abs/2111.14887 in Tab. 1.
Yes, MIC is well suitable for domain adaptation as we have shown in the paper.
from mic.
Related Issues (20)
- Varying Results across Trainings HOT 4
- Why Do Target Datasets Have ann_dir? HOT 2
- We have a question for your amazing work. HOT 1
- Pretrained model download problem HOT 3
- Memory issue HOT 1
- What is the respective file on SegFormer OneDrive for pretrained/mit_b5.pth HOT 4
- Resize problem
- Question about forward_train()
- performance variance
- Any other loss function instead of crossentropy
- Differrent GPU Memory Occupancy
- Augmentation's randomness in classification code HOT 1
- Config file for using AdvSeg as uda HOT 1
- Classification problem HOT 1
- Teacher model for inference HOT 1
- CUDA Out of memory
- Why MIC work on DeepLab with ResNet backbone?
- How to perform distributed training
- Does this code really work in the way of Unsupervised? HOT 1
- Why in custom data, the accuracy of Detectron 2 is 22% higher than MIC model?
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 mic.