search

lihong2303 / AGM

[ICCV2023] The repo for "Boosting Multi-modal Model Performance with Adaptive Gradient Modulation".
MIT License
22 stars 2 forks source link

Bugs in codes #7

Open soudelor01 opened 8 months ago

soudelor01 commented 8 months ago

Hi, I think your codes need to be carefully reviewed because there are many bugs and unused variables inside. Such as 'iteration = (epoch -1) * cfgs.batch_size + step + 1' in URFunny_main.py. The code style of scripts for different datasets is even different, making replication and migration very difficult and annoying.

lihong2303 commented 8 months ago

What is your bug? Or is there anything you do not understand? The computation of iteration refers to the step of iteration for the entire training process.

soudelor01 commented 8 months ago

What is your bug? Or is there anything you do not understand? The computation of iteration refers to the step of iteration for the entire training process.

The computation here should use len(dataloader) not batch_size. This makes the iteration as well as train_score to be totally wrong.

lihong2303 commented 8 months ago

Yes, it is a typo. I'm ready to correct it. thx!