Thanks for your interest in contributing to E2E TensorFlow Lite Tutorials!
Please follow these steps when contributing:
Step 1. (Optional) Propose a project idea with either a brief description or an overview page if you wish to add more context about it.
Step 2. (Optional) Create issues for requesting help, in general in these areas:
- Model training
- Model conversion to tflite (and add metadata) <-- tflite model can reside in a team members' GitHub, in your project repo or on TFHub. Please validate the tflite model with metadata in the Android app especially if you are publishing it to TensorFlow Hub for wider adoption.
- Android app
- iOS app
- Write an E2E tutorial
- Here is an example - How to Create a Cartoonizer with TensorFlow Lite.
Note that you can also write a blog post about model conversion processes that are generally non-trivial. Here's a good example - Optimizing MobileDet for Mobile Deployments. If you want to benchmark multiple TensorFlow Lite models and write a blog post comparing their performance aspects we welcome that as well. Refer to this blog post - A Battle of Text Detectors for Mobile Deployments: CRAFT vs. EAST if you are interested in writing such materials.
Step 3. If you are interested in helping out, please create a project repo and add the link to the E2E TFLite Tutorials README. Update project to "In Progress" section.
Step 4. After project is completed, move from "In progress " to "Completed" section on E2E TFLite Tutorials. And request with a PR to add your project to [awesome-tflite].
When createing a PR, please make sure to:
- Provide a meaningful title such as "Update README.md to [action]" and do not title it as "Update README.md".
- Make sure there are no broken links.