Google Cloud Digital LeaderData and AI with Google CloudMedium
A small startup is developing a new mobile application that uses a custom machine learning model to classify images uploaded by users. They need a cost-effective solution to train this model using their continually growing dataset of images. The solution should allow them to scale their training jobs without managing underlying infrastructure and integrate well with other Google Cloud services. Which Google Cloud AI/ML product is the most suitable for this scenario?
- AVertex AI Training
- BAutoML Vision
- CBigQuery ML
- DCloud Vision API
Show answer & explanationAnswer & explanation
Correct answer: A. Vertex AI Training
Vertex AI Training provides a managed service for running custom model training jobs at scale without managing infrastructure. It's cost-effective for custom models and integrates well with other Google Cloud services.
Why the other options are wrong
- B. AutoML Vision is for training custom models with minimal code, but for a 'custom machine learning model' and more control, Vertex AI Training is more appropriate.
- C. BigQuery ML is for training models within BigQuery using structured data, not for image classification with custom models.
- D. Cloud Vision API is a pre-trained API for image analysis, not for training custom models.
Vertex AI Training
A managed service on Google Cloud for training custom machine learning models at scale, using various frameworks.
- Supports custom training code and popular ML frameworks
- Manages infrastructure for distributed training
- Integrates with other Vertex AI services for end-to-end MLOps
Memory trick: To reach the Vertex of ML, you need strong Training.