PA Lab2 - PGPM -Quiz1

What is one way to update a deployed ML model on Render?

• Retrain the model and redeploy with the updated files

• Edit the model directly in the cloud

• Use Render’s automatic retraining feature

• Train the model with a larger dataset on Render

What is one common challenge when deploying ML models on the cloud?

• Lack of computational power

• Overfitting to training data

• Ensuring low response time (l) for real-time predictions

• Lack of labeled data

Which file format is often used to save a trained ML model for deployment?

• .csv

• .py

• .pkl or .h5

• .html

How is an ML model typically deployed as a service on Render?

• By embedding it in a mobile application

• By creating a API (app) or GraphQL API

• By integrating it with a CRM system

• By uploading it to a web browser

What is the typical cost structure for using Render platform? (apart from the limited free version)

• Flat monthly fee

• Pay-as-you-go, based on resources used

• One-time payment

• Free for unlimited usage

What is required to deploy a model API on Render?

• A trained ML model and a web framework like Flask or FastAPI

• Only the raw data used for training

• A Render mobile application

• Pre-installed cloud GPUs

Which programming languages are most commonly used to deploy ML models on Render?

• Python and JavaScript

• C++ and Ruby

• HTML and CSS

• SQL and NoSQL

What is the primary purpose of deploying an ML model?

• To collect data

• To make predictions or decisions in real-world scenarios

• To train the model on new data

• To store data securely

What is the "odd thing" here in relation to cloud deployment

• Azure

• AWS

• GCP

• Keras