Can I train with batch_size > 1

megvii-research / MOTR

[ECCV2022] MOTR: End-to-End Multiple-Object Tracking with TRansformer

Other

633 stars 93 forks source link

Can I train with batch_size > 1 #35

Closed ZihaoLiu97 closed 2 years ago

ZihaoLiu97 commented 2 years ago

Hi, thanks for your great job! But I have a small problem that whether it can be trained with batch_size > 1? But I get an error that "frame.requires_grad = False AttributeError: 'list' object has no attribute 'requires_grad'".

vaesl commented 2 years ago

Currently, the code of MOTR only support the setting of batch_size = 1.