shufan wu

[Feature][CodeCamp #67] Add Decouple Algorithm

4

## Motivation Add decouple algorithm for long-tailed recognition. ## Modification Add decouple algorithm config files and implementation files. [Paper](https://arxiv.org/abs/1910.09217) [Official Implementation](https://arxiv.org/abs/1910.09217) [Issue](https://github.com/open-mmlab/mmclassification/issues/1231) ## BC-breaking (Optional) Does the modification introduce changes...

Discriminator pretrained model

1

thank you for your work! I have implemented the train code and want to finetune in my own dataset, could you provide the discriminator pretrained model?

是否提供模型权重

2

您好！我打算在mmdetection框架下复现dense teacher，请问能提供模型权重做推理精度对齐吗？

[Feature] Support Dense Teacher

13

This pr implement dense teacher for mmdetection. paper: https://arxiv.org/abs/2207.02541v2

[Enhance] Support dumping datasets to pkl file

3

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

PaddleOCR-VL后处理合框逻辑问题

1

您好，我在看合并的实现逻辑时，发现如下几个问题，可能导致框的合并存在问题： 1. 判断框是否上下对齐的一个条件是需要与其他框存在重叠，这里该如何理解？是否少了一个not？https://github.com/PaddlePaddle/PaddleX/blob/release/3.3/paddlex/inference/pipelines/paddleocr_vl/uilts.py#L248 2. is_cross表示两个框在水平方向无重叠，且竖直方向有重叠。该逻辑合框的依据是什么？根据现在的逻辑，会将下图中3，4合并。 [yanbaor2_yanbaoPPT_6065_layout_order_res.zip](https://github.com/user-attachments/files/23470394/yanbaor2_yanbaoPPT_6065_layout_order_res.zip)