NWD

NWD copied to clipboard

where to find "supplementary" materials？

3

hi @jwwangchn where to find "supplementary" materials? since i want to use nwd to train my own dataset, so that i need to calculate the constant variable C. thanks.

zimenglan-sysu-512

MMCV version mismatch

4

The [mmcv-nwd](https://github.com/jwwangchn/mmcv-nwd.git) provided is 1.2.7 while this repo requires 1.3.0+.

Sharpiless

Can it be implemented on TOOD?

1

it can be implemented on TOOD?The name of your method in the code is BboxDistanceMetric?

okt7760

Function wasserstein_nms_op not implemented

2

The question is I found the wasserstein_nms_op function is not implemented, which only leaves a pass.  (The source code is presented [here](https://github.com/jwwangchn/NWD/blob/main/mmdet/core/post_processing/wasserstein_nms.py#L10)) This function is used to compute NMS...

lzk901372

Can you provide a time estimate of when the benchmark and trained models will be publicly available?

ale93111

请问Wasserstein Loss可以应用于图像分割吗？如果有这方面的实验方便提供一下代码吗？

2

DINGdef

模型使用问题

1

rpn_proposal=dict( nms_pre=3000, max_per_img=3000, nms=dict(type='wasserstein_nms', iou_threshold=0.85)), 请问使用coco数据集如何调整阈值，0.85相当于nms阈值多少呢

kming123456

你好，请问config文件中提到的_aitod_training_v1.0.json_在哪里？ 我把complete_annotations解压后，没有找到_aitod_training_v1.0.json_。 谢谢，期待您的回复。

Cathy1900

When will you provide the Inference code and pretrained Models?

zhouyizhuo