Inconsistant configuration between the repo and the paper

Question

Inconsistant configuration between the repo and the paper

FrontierBreaker opened this issue 2 years ago · 3 comments

Hi thanks for your wonderful jobs! However, I notice that the config for the semantic segmentation task is not consistent with the paper, thus I met with a failure in reproducing the numbers.

Could you please provide the configs that are aligned with the configs in the main paper to foster the further research.

Thanks~

Answer 1 · 2023-03-09T09:34:41.000Z

Hi, currently we have only released the checkpoint and the evaluation scripts. The training configuration will be released soon.

Answer 2 · 2023-03-09T09:44:22.000Z

This is just the config for evaluation. There are some differences between training/evaluation configs (lr and optimizer config)

Answer 3 · 2023-03-09T09:50:19.000Z

Got that, thanks~ Looking forward to the code release