Flask application that can predict on any input data on specified model from a scikit-learn. Predictions for a given input is made from the latest trained checkpoint stored We can also train/retrain a new model and store it for later use.

Any sklearn model can be used for prediction.

pip install -r requirements.txt

 [
  {'cement': 540,  'slag':0.0 ,  'ash':0.0 , 'water':162.0 , 'superplastic':2.5 , 'coarseagg':1040.0  ,'fineagg':676.0  ,'age':28 }
 ]

{'prediction': [ 53.463463286890814]}

Run main.py and open 0.0.0.0:5000/ in the browser and use any of the options below

• /train (GET) - Trains the model.

• /predict (GET) - The Predictions are returned in json format.

• /clearModel (GET) - Deletes all the model stored in the models folder.

• Train.
• clear.
• Predict.
• Hardcoded testing
• Save checkpoint for individual(any number of) models.
• UI for input, output and to download checkpoint.