Inconsistent `from_logits` parameter across loss functions

[AndrewFalkowski]

DiceLoss has a from_logits parameter to handle both logits and probabilities, but FocalLoss does not. This creates an inconsistency when using models with softmax activation and requires awkward workarounds.

model = smp.create_model(..., activation='softmax')
outputs = model(x)  # Probabilities [0, 1]

# This works
dice_loss = smp.losses.DiceLoss(mode='multiclass', from_logits=False)

# This fails - FocalLoss applies softmax to output probabilities resulting in incorrect loss calc
focal_loss = smp.losses.FocalLoss(mode='multiclass')

Currently need to either:

1. Remove model activation and use logits everywhere
2. Manually wrap FocalLoss to convert probabilities back to logits

FocalLoss should have a from_logits parameter like DiceLoss for consistent API.

Environment

• segmentation-models-pytorch version: 0.5.0
• PyTorch version: 2.7.1

[qubvel]

Thanks for the issue! Would you like to make a fix? I would appreciate the help

[AndrewFalkowski]

Yeah, I can take a stab at it!