shuangli59/person-search-with-natural-language-description

Person Search with Natural Language Description

person-search-with-natural-language-description's Issues

Which dataSet did you use?

# as you mentioned:
"Request the dataset from sli [at] ee.cuhk.edu.hk or xiaotong [at] ee.cuhk.edu.hk (academic only)."
# Person Re-identification Datasets include lots of dataset including:

• VIPeR | 2007 | 632 | 2 | 1264 | Hand | 128X48 |   |   |  
• ETH1,2,3 | 2007 | 85,35,28 | 1 | 8580 | Hand | Vary | ✔ | ✔ | ✔
• QMUL iLIDS | 2009 | 119 | 2 | 476 | Hand | Vary | ✔ |   |  
• GRID | 2009 | 1025 | 8 | 1275 | Hand | Vary |   |   |  
• CAVIAR4ReID | 2011 | 72 | 2 | 1220 | Hand | Vary | ✔ |   |  
• 3DPeS | 2011 | 192 | 8 | 1011 | Hand | Vary | ✔ |   | ✔*
• PRID2011 | 2011 | 934 | 2 | 24541 | Hand | 128X64 | ✔ | ✔ | ✔*
• V47 | 2011 | 47 | 2 | 752 | Hand | Vary | ✔ |   | ✔
• WARD | 2012 | 70 | 3 | 4786 | Hand | 128X48 | ✔ | ✔ |  
• SAIVT-Softbio | 2012 | 152 | 8 | 64472 | Hand | Vary | ✔ | ✔ | ✔
• CUHK01 | 2012 | 971 | 2 | 3884 | Hand | 160X60 | ✔ |   |  
• CUHK02 | 2013 | 1816 | 10(5 pairs) | 7264 | Hand | 160X60 | ✔ |   |  
• CUHK03 | 2014 | 1467 | 10(5 pairs) | 13164 | Hand/DPM | Vary | ✔ |   |  
• RAiD | 2014 | 43 | 4 | 6920 | Hand | 128X64 | ✔ |   |  
• iLIDS-VID | 2014 | 300 | 2 | 42495 | Hand | Vary | ✔ | ✔ |  
• MPR Drone | 2014 | 84 | 1 |   | Pyramid Features(ACF) | Vary | ✔ |   | ✔
• HDA Person Dataset | 2014 | 53 | 13 | 2976 | Hand/Pyramid Features(ACF) | Vary | ✔ | ✔ | ✔
• Shinpuhkan Dataset | 2014 | 24 | 16 |   | Hand | 128X48 | ✔ | ✔ |  
• CASIA Gait Database B | 2015(*see below) | 124 | 11 |   | Background subtraction | Vary | ✔ | ✔ | ✔
• Market1501 | 2015 | 1501 | 6 | 32217 | Hand/DPM | 128X64 | ✔ |   |  
• PRW | 2016 | 932 | 6 | 34304 | Hand | vary | ✔ |   | ✔
• Large scale person search | 2016 | 11934s | - | 34574 | Hand | vary |   |   | ✔
• MARS | 2016 | 1261 | 6 | 1191003 | DPM+GMMCP | 256X128 | ✔ | ✔ |  
• DukeMTMC-reID | 2017 | 1812 | 8 | 36441 | Hand | Vary | ✔ |   | ✔
• DukeMTMC4ReID | 2017 | 1852 | 8 | 46261 | Doppia | Vary | ✔ |   | ✔
• Airport | 2017 | 9651 | 6 | 39902 | ACF | 128X64 | ✔ |  

training loss

I re-implement this model with Pytorch, but the loss is around 0.69 and does not drop in the training stage.

I want to know the loss when you training.

Thanks~

Can I run your code without GPU?

数据集

您好，请问这个数据集如何去获取？ 找了很长时间，也没找到开源的数据集，谢谢了 是需要邮箱吗？ [email protected]
我记得以前用邮箱联系过您，寻求数据集，可是一直没回复。

test_query/p2703_s3526.jpg has 10 captions.

Model not found in snapshot folder- lstm1_rnn512_bestACC.t7

can you please share the snapshot folder which contains snapshot/lstm1_rnn512_bestACC.t7 to run Retrieval.lua file. I am unable to find the file under snapshot folder shared by you in your github.

the sentence descriptions are not 80412?

hello, thank you very much for providing your dataset. I have a question about the dataset. In this paper, it says there are 40206 images and 80412 sentence descriptions, however the result of my experiment shows that there are 40206 images and 80440 sentence descriptions. If i compare with your work, how i to deal with dataset? Should i use all dataset or remove the extra sentences? Thans.

Problem with model loading -- I am using LUA5.1. Could you please let me know the torch version?

Found Environment variable CUDNN_PATH = /home/iiitd/cuda/lib64/libcudnn.so.5/data/home/kajalk/torch/install/bin/luajit: ...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:351: read error: read 709434 blocks instead of 134260544 at /data/home/kajalk/torch/pkg/torch/lib/TH/THDiskFile.c:356
stack traceback:
[C]: in function 'read'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:351: in function <...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:245>
[C]: in function 'read'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:351: in function 'readObject'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
/data/home/kajalk/torch/install/share/lua/5.1/nn/Module.lua:192: in function 'read'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:351: in function 'readObject'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
/data/home/kajalk/torch/install/share/lua/5.1/nn/Module.lua:192: in function 'read'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:351: in function 'readObject'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
...a/home/kajalk/torch/install/share/lua/5.1/torch/File.lua:409: in function 'load'
Retrieval.lua:52: in main chunk
[C]: in function 'dofile'
...jalk/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
[C]: at 0x00405d50

请问CUHK-PEDES数据集怎么下载的？发邮件也没用呀？

请问CUHK-PEDES数据集怎么下载的？发邮件也没用呀？

how can I get the dataset?

Thanks for sharing such a good work. I want to know how I can get the access to the data? still send an email to these two mail address?

questions about train parameters

Thank you for your public code.