search

SLR1999 / Human-Aware-Motion-Deblurring

This repo is the implementation of the paper Human-Aware Motion Deblurring (https://arxiv.org/abs/2001.06816)
5 stars 1 forks source link

Question to attention network pretraining #1

Open GangMieYe opened 3 years ago

GangMieYe commented 3 years ago

Thx for your work.The paper mentioned that attention network is pretrained 70000 iteration for convergence,could you please tell me how to do that?

SLR1999 commented 3 years ago

you can change the default number of epochs in train.py or required number of epochs can be passed as a command line argument

GangMieYe commented 3 years ago

Thx for your kind reply.But in train.py attention network and the whole deblurring network are trained together,there is no setting for attention network pretraining.Did I ingore that?

SLR1999 commented 3 years ago

Our code just trains them together. With the current implementation it is not possible, but it can be achieved by writing a separate data loader and train file