CameraTraps/roadmaps.md

Development roadmap

Here you can find details of the milestone roadmap for Pytorch-Wildlife:

October 0.1

✔ Packaging

• Add the YOLOv5 detection model.
  
• Add the Megadetector pretrained weights.
  
• Add a Notebook to create, load and run inference on the YOLOv5 detection model using Megadetector pretrained weights.
  
• Add an animal species classification model and pretrained weights.

November 0.2

✔ Packaging

• Add a documented animal video detection pipeline.
  
• Add a user-friendly interface for animal detection.
  
• Add a module for custom datasets.
  
• Add a Jupyter notebook for the video detection tutorial.

December 0.3

✔ Packaging

• Add the video detection pipeline to the user interface.
  
• Add animal classification to the user interface.
  
• Complete documentation on how to use the user-interface.
  
• Add a module for fine-tuning and training from scratch.
  

January 0.4

✔ Packaging

• Add a utilities module for data pre-processing and post-processing.
  
• Add Snapshot Serengeti classification weights.