search

huanghoujing / person-reid-triplet-loss-baseline

Rank-1 89% (Single Query) on Market1501 with raw triplet loss, In Defense of the Triplet Loss for Person Re-Identification, using Pytorch
487 stars 129 forks source link

Other CNN Models #1

Closed ahkarami closed 6 years ago

ahkarami commented 6 years ago

Dear @huanghoujing, Have you ever trained other ConvNets (e.g., ResNet101, ResNet152, ResNext, Inception-ResNet-V2, DenseNet) as the backbone of your code?

huanghoujing commented 6 years ago

Hi, I only tried ResNet-50. It's easy to replace the model with yours in file tri_loss/model/Model.py. Your model has to return feature with shape [batch_size, out_channel] to make it consistent with the training script.

ahkarami commented 6 years ago

Dear @huanghoujing, Thank you very much for your answer.

huanghoujing commented 6 years ago

Thanks for attention : )