Flask API for model serving in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of this Flask API response code?

Consider a Flask API that loads a trained model and predicts the class for input data. What will be the JSON response when sending input data [5.1, 3.5, 1.4, 0.2] to this endpoint?

ML Python

from flask import Flask, request, jsonify
import numpy as np
import pickle

app = Flask(__name__)

# Load model
with open('iris_model.pkl', 'rb') as f:
    model = pickle.load(f)

@app.route('/predict', methods=['POST'])
def predict():
    data = request.json['data']
    arr = np.array(data).reshape(1, -1)
    pred = model.predict(arr)
    return jsonify({'prediction': int(pred[0])})

A {"prediction": 0}

B {"prediction": "setosa"}

C {"prediction": 1}

D {"error": "Invalid input"}

💡 Hint

The model is trained on Iris dataset where class 0 corresponds to 'setosa'. The input matches typical setosa features.

The model predicts class 0 for the given input features which correspond to Iris setosa. The API returns the class index as an integer in JSON.

❓ Model Choice

intermediate

Which model type is best suited for this Flask API serving handwritten digit recognition?

You want to serve a model via Flask API that recognizes handwritten digits (0-9) from 28x28 grayscale images. Which model type below is most appropriate for this task?

A A k-means clustering model

B A linear regression model

C A decision tree classifier

D A convolutional neural network (CNN)

💡 Hint

Consider the input type and the need to capture spatial features in images.

CNNs are designed to capture spatial patterns in images and perform well on image classification tasks like digit recognition.

❓ Hyperparameter

advanced

Which hyperparameter tuning improves Flask API model latency the most?

You deployed a deep learning model in a Flask API. The API latency is high. Which hyperparameter tuning below will most likely reduce prediction latency?

A Increase the number of layers in the model

B Reduce model size by pruning or quantization

C Increase batch size during training

D Increase learning rate during training

💡 Hint

Think about what affects model size and inference speed.

Reducing model size via pruning or quantization decreases computation needed during inference, lowering latency.

🔧 Debug

advanced

Why does this Flask API code raise a KeyError?

This Flask API code raises a KeyError when a POST request is sent. What is the cause?

ML Python

from flask import Flask, request, jsonify
app = Flask(__name__)

@app.route('/predict', methods=['POST'])
def predict():
    data = request.json['input']
    # model prediction code here
    return jsonify({'result': 'ok'})

A The route should accept GET instead of POST

B Flask app is missing debug=True

C The JSON payload does not contain the key 'input'

D The jsonify function is used incorrectly

💡 Hint

Check the JSON keys sent in the request body.

If the client sends JSON without the 'input' key, accessing request.json['input'] raises KeyError.

❓ Metrics

expert

Which metric best evaluates a Flask API model for imbalanced binary classification?

Your Flask API serves a model that predicts rare disease presence (positive class is 1% of data). Which metric below is best to evaluate model performance?

A F1-score

B Precision

C Recall

D Accuracy

💡 Hint

Consider a metric that balances false positives and false negatives in imbalanced data.

F1-score balances precision and recall, providing a better measure than accuracy alone for imbalanced classes.