Use 2016_person_re-ID train DukeMTMC

layumi / 2016_person_re-ID

TOMM2017 A Discriminatively Learned CNN Embedding for Person Re-identification

https://dl.acm.org/citation.cfm?id=3159171

MIT License

265 stars 116 forks source link

Use 2016_person_re-ID train DukeMTMC #14

Open hsqaihkl opened 6 years ago

hsqaihkl commented 6 years ago

Hi, Zhedong Zheng; I used the master to train Duke Data , but the mAP and the r1 is too low. I just change faster.m and crazy.m the image path.

layumi commented 6 years ago

Hi @hsqaihkl Did you change the training code for dukemtmc-reid? Or you just used the model trained on Market-1501?

hsqaihkl commented 6 years ago

I guessed the reason was used the Market-1501（resnet52_2stream_drop0.9/net-epoch-75.mat).And now I use prepare_data.m and train_id_net_res_2stream.m to train it ,but I don't know is that right ways.

hsqaihkl commented 6 years ago

@layumi I'm sorry，I'm a newbie in deeplearning

layumi commented 6 years ago

No worries @hsqaihkl You may check your code with my baseline code. https://github.com/layumi/DukeMTMC-reID_baseline

hsqaihkl commented 6 years ago

OK，Thks