Microsoft Azure Custom Vision Client Library for Python (version 3.1.1). Allows building, training, and deploying custom image classifiers and object detection models. Low-level functionality; newer azure-ai-vision SDK is recommended. This package is in maintenance mode: no new features, only critical bug fixes. Release cadence is irregular.
pip install azure-cognitiveservices-vision-customvisionNo compatibility data collected yet for this library.
Verified import paths — ran on the pinned version, not inferred.
Initialize training client and create a project. Set TRAINING_ENDPOINT and TRAINING_KEY env vars.
Use azure-ai-vision for new projects; see https://aka.ms/azsdk/azure-ai-vision
Always use ApiKeyCredentials(in_headers={'Training-key': key}) for training API.Obtain the prediction endpoint from the Custom Vision portal (e.g., https://<region>.api.cognitive.microsoft.com). For prediction, use PredictionEndpoint from the portal settings.
Use correct import: from azure.cognitiveservices.vision.customvision.training import CustomVisionTrainingClient
Use ApiKeyCredentials(in_headers={'Training-key': key}) for training. For prediction, use ApiKeyCredentials(in_headers={'Prediction-key': key}).Ensure you pass ApiKeyCredentials with either 'Training-key' or 'Prediction-key' header.