Courseiva

AI-900 Practice Question: Describe features of computer vision workloads on Azure

What is 'model export' in Azure Custom Vision and what formats are supported?

⚠ Common exam trap

It's easy for candidates to confuse 'model export' with exporting training data or logs, because Azure Custom Vision does offer data export options elsewhere, but the specific term 'model export' refers exclusively to the trained model artifact for offline deployment.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

Exporting trained models as ONNX, TensorFlow, CoreML, or Docker for offline/edge deployment

Model export in Azure Custom Vision allows you to export a trained image classification or object detection model in formats like ONNX, TensorFlow, CoreML, or Docker container images. This enables the model to run offline on edge devices or local servers without requiring a continuous connection to the Azure cloud, which is critical for low-latency or disconnected scenarios.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

  • Exporting model training logs and metrics to Excel for analysis

    Why it's wrong here

    Exporting model training logs and metrics to Excel is a reporting action, not model export — metrics like precision, recall, and loss curves describe training quality, but these numbers are not the model itself. Exporting them to Excel supports analysis and auditing, not deployment; model export in Custom Vision packages the learned weights for runtime inference, not the performance history.

  • Exporting trained models as ONNX, TensorFlow, CoreML, or Docker for offline/edge deployment

    Why this is correct

    Exporting trained models as ONNX, TensorFlow, CoreML, or Docker is model export because it packages the learned weights and inference graph into formats that run on edge/offline devices — ONNX is cross-platform, TensorFlow targets Android/Linux, CoreML targets Apple devices, and Docker containers expose the model as a local REST API. This allows inference without cloud calls, meeting latency, privacy, or connectivity constraints.

  • Exporting the training data to another Azure service for fine-tuning

    Why it's wrong here

    Exporting the training data to another Azure service for fine-tuning moves raw images and annotations for further training, not the trained model's weights — the export is about dataset portability for iterative development, whereas model export is the deployment artifact. Fine-tuning a separate model in another service is a data-management workflow; it doesn't deliver the original Custom Vision model to an offline inference endpoint.

  • Exporting a Custom Vision project as a YAML configuration file for source control

    Why it's wrong here

    Exporting a Custom Vision project as a YAML configuration file for source control is not model export — it serializes project metadata such as tags, training iterations, and hyperparameters for versioning, but it does not package the trained model weights into an inference-ready artifact. A YAML config cannot be consumed by an edge device to run predictions; model export specifically produces executable model files or containers for offline inference.

About these practice questions

This AI-900 question is part of Courseiva's 985-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →

How Courseiva writes practice questions · Editorial policy

JA

Written by Johnson Ajibi, MSc IT Security

Senior Network & Security Engineer · founder of Courseiva

This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.