Lukas Hoyer

Thanks for your interest in our work. As explained in the comment of the referenced code `self.source[i1]` samples a new crop from i1. The cropping is included in the data...

If you comment out these lines, the image is chosen to contain a rare class. However, there is no guarantee that the random crop of the image will contain this...

Hi @kimkj38, I did not evaluate the teacher performance for MIC. However, in previous UDA studies, we found that the student and teacher have similar performance at the end of...

Dear @kaigelee, The results will vary even with the same code and seed as pytorch has multiple sources of nondeterministic behavior that cannot be completely eliminated (e.g. differentiating bilinear upsampling)...

To ensure meaningful ablations considering the influence of randomness, we reported the mean over three training runs.

These parameters are for a feature that I have discarded later. In the provided config, this feature is disabled anyway. After removing the related code, I ran MIC again and...

Hi @baek85, Thank you for your interest in our work! Yes, your are right. This is indeed an unintended behavior. However, to ensure consistency of the published results and provided...

You can use https://github.com/lhoyer/MIC/blob/67fe7dd3e81875bee6f07f4f90567ed5b71294ba/seg/experiments.py#L417 to generate a config file for advseg. Please, also have a look at https://github.com/lhoyer/MIC/tree/master/seg#training on additional information how to use experiments.py.

Thank you for your interest in our work! It is hard to guess what is going wrong without knowing details about the used source and target datasets. I would recommend...

Hi @jiyeooong, Thank you for your interest in our work! To use HRDA in a domain generalization setting (source-only), please refer to the following instructions: https://github.com/lhoyer/HRDA?tab=readme-ov-file#domain-generalization Best, Lukas