Comments (2)
Dwrety
commented on August 28, 2024
1
Interesting. Thanks for your answer, I reimplemented your OneNet in MMdet framework and got a better result with 34+ mAP in mask. After bring mask cost into the equation, I was able to get around 35.6 mAP in mask. I think the mask branch learns well enough, the cap in performance is mostly because of classification error, and this could be because of underfitting with only 1 anchor for each GT.
from onenet.
PeizeSun
commented on August 28, 2024
Hi~
- OneSeg is just a naive combination of OneNet and CondInst, we don't dig out details.
- I guess the assignment results from matching is a little different from hand-crafted methods. For example in human detection(figure9 in our paper), the assignment results of positive samples are inside heads, but the hand-crafted positive samples are in the body parts.
from onenet.
Related Issues (20)
- util/box_ops.py 尝试过diou，ciou、？ HOT 1
- Hi. Have you ever tried using SGD? HOT 1
- why the large and medium object‘Ap is lower than the centernet‘s
- Pretrained weight and Sparse RCNN implementation HOT 2
- Cost and loss difference HOT 2
- What differences between min cost assignment in OneNet and Hungarian algorithm in Sparse R-CNN? HOT 2
- KeyError: 'Non-existent config key: MODEL.OneNet' HOT 1
- Setting of learning rate HOT 2
- The loss is Nan HOT 3
- Hi!, when num_classes ==1, cant compute focal loss, because if p.shape[-1] ==self.num_classes: nn.init.constant_(p, self.bias_value) HOT 1
- It seems that there is a typo in table 5 . HOT 2
- Why is my accuracy at 1.441？ HOT 2
- DLA34 Backbone Result HOT 1
- Questions about focal loss used, with comparison to CenterNet HOT 4
- Why change the matcher from Minimum Cost to Hungarian? HOT 2
- 请问，有MMDetection实现吗？ HOT 2
- Ablation on original training losses? HOT 3
- where are the multi-head training implement? HOT 2
- about the index of positive
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 onenet.