qubvel-org/segmentation_models.pytorch

Add segmentation + classification head fine-tuning example

Open

#912 opened on Aug 23, 2024

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (1,835 forks)github user discovery
examplesgood first issue

Repository metrics

Stars
 (11,601 stars)
PR merge metrics
 (PR metrics pending)

Description

It would be great to have more examples for fine-tuning in the library!

Similar to current examples for binary-segmentaion and multi-label segmentation would be great to have example for segmentation + classification head. The same camvid or oxford pet dataset can be used. As a classification target for these datasets, the "existence" of the class on the image can be used.

The example should showcase how to

  • fine-tune a model with pytorch-lightning (or any other fine-tuning framework, even a plain pytorch)
  • use loss function for segmentation and classification
  • compute metrics for classification and segmentation heads
  • visualize results

Docs on how to create a model with classification head.

In case anyone wants to work on this you are welcome! Just notify in this issue, and then ping me to review a PR when it's ready.

Contributor guide

Research direction
Explore the segmentation models.pytorch library to create a model with an auxiliary classification head. Use the provided documentation on creating such a model. Then, implement a fine tuning script using PyTorch Lightning (or plain PyTorch) that combines segmentation loss and classification loss, computes metrics for both tasks, and visualizes results. Use a dataset like CamVid or Oxford PET, where classification target is the existence of a class in the image.
Tech stack
pythonpytorch
Domain
machine learningbackend
Issue type
Documentation
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PythonPyTorchsegmentation models.pytorch
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
60