Anvil Banana

Here we build a server for an application that can be used to check the expiry date of unpeeled bananas.

The UX runs on anvil (https://vjkvak3xbkikrszw.anvil.app/E5OCO454257ZXDTORU6EC7ME) and the model was built in Google Colab as part of an ML training hosted by #UnpackAI

Food spoilage and waste are responsible for 30% of production loss. The application built here is a simple prototype to explore how AI can be applied to reduce the waste of fresh products.

See https://github.com/mmrmas for more (upcoming) projects

Install the necessary packages and modules

!pip install -Uqq fastbook !pip install anvil-uplink

from fastbook import * from fastai.vision.widgets import * from pathlib import Path import anvil.server

Connect to the Anvil app. For security reasons, don't publish output from this server in the application directly, but instead generate the outcome in the terminal from.

anvil.server.connect("XMVM2EQVXJEOC3RJYKQPZRHZ-VJKVAK3XBKIKRSZW")

Run the prediction and send back the classification. The classification is translated into a standardized output string in the terminal

import anvil.media #get the pre-trained model !wget -cq https://www.dropbox.com/s/d3hhi2118p2l8b3/bananaModel.pkl?dl=0 learner = load_learner('/work/bananaModel.pkl?dl=0') @anvil.server.callable def predict_banana(file): with anvil.media.TempFile(file) as f: img = PILImage.create(f) pred, pred_idx, probs = learner.predict(img) prediction = pred return prediction