Francesco Marzola

Unstable performances in validation set and prediction collapse

9

Hi everyone, many thanks to @LeeJunHyun for the awesome implementations. I'm using this repo to segment medical grayscale images on a private dataset. I have a problem in the training...

Adapting code to mimic scipy.interpolate.griddata

1

Hi @sbarratt, thanks for this implementation. I tested it against scipy.interpolate.griddata and I got very close results using torch and 'linear' interpolation. Have you noticed these differences? Are you planning...

Losses during training

3

Hi all, congratulations on your work! I'm trying to use the repo for one of my projects. During training, losses stabilize very quickly in about 100k steps. Is it something...