Add "The One Hundred Layers Tiramisu" model
Opened this issue · 1 comments
I implemented the "The One Hundred Layers Tiramisu" model as a part of Google Code-In. The code can be viewed in this Google Colab.
Because there is no image segmentation model available as a part of this repository yet, I would like to contribute by adding the model. I have added another model to another TensorFlow repository, which I also built as a part of Google Code In. Adding the repository is not a part of the task, but I would still like to do it because it benefits the community.
I don't think pre-trained weights are available for this particular model. Considering other models in this repository do provide pre trained weights, will that be a problem? In the worst case we can train the model ourselves and use those weights.
I'm not sure how adding a model works in this repository, since CONTRIBUTING.md is empty. Can I fork this repository, add the model, and create a PR?
Has anyone had time to look into this yet?