开源数据集

[whz-snn]

请问有没有可以wget的链接直接获取数据集，220G百度网盘太慢了

[wangxiao5791509]

@whz-snn 目前还没有。可以考虑开个会员。

[linyuming0722]

按照README里面的指示配置环境之后出现了‘cannot import name 'Config' from 'mmcv'这个错误，一开始我认为是mmcv版本的问题，但是换了很多版本尝试过了，mmcv2.0以上的版本都是报‘cannot import name 'Config' from 'mmcv’这个错误，1.4-1.7版本的会报‘cannot import name 'CUDA_HOME' from 'mmcv.utils'’错误，请问要怎么解决呢？

[wangxiao5791509]

@WuZongzhen

[QiWang233]

Hi, Wang:

1. When I try to recurrent, I found these missing parameters in the above picture.I wonder if you can tell me the meaning of these parameters in detail.

2. Also, I got three list files ('test_label' 'train_label' and 'val_label') in Baidu Netdisk which I don't know how to use it correctly.

Looking forward to your reply.

Best wishes Qi

[WuZongzhen]

按照README里面的指示配置环境之后出现了‘cannot import name 'Config' from 'mmcv'这个错误，一开始我认为是mmcv版本的问题，但是换了很多版本尝试过了，mmcv2.0以上的版本都是报‘cannot import name 'Config' from 'mmcv’这个错误，1.4-1.7版本的会报‘cannot import name 'CUDA_HOME' from 'mmcv.utils'’错误，请问要怎么解决呢？

You can try running the code in the following version: cuda 9.1 , torch 1.8.1+cu111, mmcv-full 1.7.0

[WuZongzhen]

1. When I try to recurrent, I found these missing parameters in the above picture.I wonder if you can tell me the meaning of these parameters in detail.
2. Also, I got three list files ('test_label' 'train_label' and 'val_label') in Baidu Netdisk which I don't know how to use it correctly.

This section is the dataset path, you can check the updated files.