This repo classifies the input butterfly image among the 10 classes of butterfly. The model is created using teachable machine
Dataset Used for training model: Leeds Butterfly Dataset
Input: Image or image path
Output: predicted label and prediction probability
Labels:
Danaus plexippus
Heliconius charitonius
Heliconius erato
Junonia coenia
Lycaena phlaeas
Nymphalis antiopa
Papilio cresphontes
Pieris rapae
Vanessa atalanta
Vanessa cardui
To Run:
- Install tensorflow servering
$ sudo apt-get update
$ sudo apt-get install tensorflow-model-server
$ tensorflow_model_server --version
TensorFlow ModelServer: 2.2.0-rc2+dev.sha.d22fc19
TensorFlow Library: 2.2.0
- Starting tensorflow servering server
$ tensorflow_model_server --model_base_path=path_to_model/model/ (or absolute path for the model) --rest_api_port=9000 --model_name=butterfly_classify(or name of your serving server)
- Start flask server
python app/app.py
- Use postman/insomnia's post request, input image as filetype and get the prediction
http://127.0.0.1:5000/predict/butterfly/
OR
Upload the image file and click predict button at http://127.0.0.1:5000/
