-
notebookで実行中
-
Hi ,
I am facing a key error while pseudo labeling. Looks like pos_pid selected is not found in the corpus.
```python
INFO [gpl.toolkit.pl.run:60] Begin pseudo labeling
.....
File ~gpl/toolk…
-
### Search before asking
- [X] I have searched the YOLOv8 [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussions) and fou…
-
As I looked at your code, "model.eval()" is not used. (Instead, "torch.no_grad()" is used)
( ex.
https://github.com/charlesCXK/TorchSemiSeg/blob/main/exp.voc/voc8.res50v3%2B.CPS%2BCutMix/train.py
…
-
## 세부 설명
test set에 지금까지의 best submit csv파일로 labeling을 한 후, train set에 test 파일도 같이 넣어 훈련을 해보자.
## 해야 하는 것
* [x] best.csv 파일을 coco-format의 json 파일로 변환
* [ ] pseudo labeling한 json파일을 가지고 다시 train을 …
-
Hi, I read your paper and would like to try your approach on my data.
However, I realized the `pseudo_labeling` function is missing. The file `pseudo_labeling.py` seems to be empty. Would you be abl…
-
请问使用该策略在训练早期会不会造成大量信息损失呢?因为初始阶段分类器未曾训好,导致两种伪标签的一致性很差,只有很少的像素被纳入训练?
请问您考虑过这个问题吗?
-
您好,感谢您卓越的工作,我认真阅读了您的代码,有了以下的理解:
1. 关于Mix pseudo-labeling:
a. 用当前批处理的每个像素点的特征与每个类的原型的特征相似度作为一类标签,另一种就是常规的经过网络预测之后分类的标签。用两者的交集来监督网络的训练。
2. 关于 Cross pseudo-labeling:
a. 先使用Logit Space训练…
-
After reading your article, I have some questions and would like to seek your clarification. My questions are as follows.:
1. Were the SM3 pre-training weights used in your paper obtained through pre…
-
In your paper, you present a comparison with Full Labels (Fig 1. right)
What does "full" mean in the case of LaserMix? Since it is a semi-supervised approach, what does training with the full train s…