Mask loss with soft labels

facebookresearch / Mask2Former

Code release for "Masked-attention Mask Transformer for Universal Image Segmentation"

MIT License

2.59k stars 388 forks source link

Mask loss with soft labels #236

Open antopost opened 7 months ago

antopost commented 7 months ago

Hey @bowenc0221 I was wondering if in theory it would be possible to train a model that uses mask loss with soft label ground truth. If so, would any alterations have to be done to the code or could I pass the prediction and soft labels to the loss function as is?