bermanmaxim/LovaszSoftmax
A loss function layer implementing the Lovász-Softmax surrogate for optimizing IoU in neural network training.
Not currently ranked — collecting fresh signals.
star history
The repository provides PyTorch and TensorFlow implementations of the Lovász-Softmax loss, a differentiable surrogate for the intersection-over-union metric used in image segmentation. It includes binary and multiclass variants suitable for training neural networks on semantic segmentation tasks. Published at CVPR 2018.
Frequently asked
- What is bermanmaxim/LovaszSoftmax?
- A loss function layer implementing the Lovász-Softmax surrogate for optimizing IoU in neural network training.
- Is LovaszSoftmax open source?
- Yes — bermanmaxim/LovaszSoftmax is open source, released under the MIT license.
- What language is LovaszSoftmax written in?
- bermanmaxim/LovaszSoftmax is primarily written in Jupyter Notebook.
- How popular is LovaszSoftmax?
- bermanmaxim/LovaszSoftmax has 1.4k stars on GitHub.
- Where can I find LovaszSoftmax?
- bermanmaxim/LovaszSoftmax is on GitHub at https://github.com/bermanmaxim/LovaszSoftmax.