I'm getting weird colored artifacts after running the sample training with vimeo90k triplets dataset. All the arguments given to the train.py were exactly the same of the README.md. This does not happen with the downloaded pretrained model checkpoint. Any idea of what could be causing this problem?
I'm getting weird colored artifacts after running the sample training with vimeo90k triplets dataset. All the arguments given to the train.py were exactly the same of the README.md. This does not happen with the downloaded pretrained model checkpoint. Any idea of what could be causing this problem?