x-brainlab / pi-reid Goto Github PK

Where can i get the file cuhk_pre_index.json?
Can you share it ?

Hello, I sent an email to your Google Email to ask for the dataset three days ago, but I haven't received the reply yet. Could you please check your email? Thanks

您好，我需要数据集进行研究，三天前已经给您的谷歌邮箱发了邮件，还没有收到回复，可以
麻烦您查看一下邮箱么。

要如何才能联系到作者拿到数据库下载链接

Thanks for your helpful work!
When I try to reproduce the result in paper, I find the pretrained model download from Baidu disk can't work well
By the way, what's your torch version? Both 1.6 and 1.5 do not work for me.
Error message using pytorch 1.6：
“”“
torch.nn.modules.module.ModuleAttributeError: 'BatchNorm2d' object has no attribute '_non_persistent_buffers_set'
”“”
Error message using pytorch 1.5：
“”“
AttributeError: module 'torch.jit' has no attribute '_script_if_tracing'
python-BaseException
”“”
looking forward to your reply : -)

Thanks for you great work! Cloud you share you PI-PRW and PI-CUHK-SYSU datasets with me? Thanks. My email address is [email protected].

Thanks for you great work! I didn’t find the test code. Did I miss something?

Hi, I want to traning and test this project, but I find the open source dateset prw unusable, please share the new dataset pi-prw address to this email [email protected], thankyou.

Hello everyone,

I have a question to your code: In line 204 of pisnet in file PI-ReID/modeling/backbones/pisnet.py the forward pass returns the following during training:
return x1, attention1, x2, attention2, x_g, x3, attention3
with x_g being the gallery features. However in line 70 in file PI-ReID/modeling/PISNet.py you name the previous outputs as follows:
feature_gallery, gallery_attention, feature_gallery1, gallery_attention1, feature_query, reg_feature_query, reg_query_attention
The third last output, i.e., the gallery features x_g in pisnet.py is used as feature_query in the second file.

Is that a bug or am I missing something?

Best and thanks in advance :)
Jenny