This repo is strongly based on https://github.com/microsoft/onnxruntime-inference-examples/tree/main/mobile/examples/image_classification/android. Follow their steps to build and run the app.
The differences between both repositories are:
- You can pick an image from your gallery in this implementation;
- Quantization option removed.
To use your classifier, you must:
- convert your model to ORT format, rename it to "model.ort" and paste it to
app/src/main/res/raw/
; - create a txt file named "labels.txt" with your labels at each line at the same order of the prediction index and paste it to
app/src/main/res/raw/
;
You can change the name of your App at app/src/main/res/values/strings.xml
.
In order to use it in your phone, select Build -> Build Bundle(s) / APK(s) > Build APK(s).
The apk file will be located in project-name/module-name/build/outputs/apk/
.
Example screenshot of a bird species predictor app trained with https://www.kaggle.com/gpiosenka/100-bird-species.