Repository navigation
🌟 Summary
Version 0.1.90 updates the Ultralytics Platform API and Python SDK with more flexible image prediction and similar-image search.
📊 Key Changes
- 🖼️ Visual prompts for image prediction: The sync and async SDK methods now accept up to 200 same-image seed labels—boxes, segments, or oriented boxes—for YOLOE26X prediction. Predictions can include confidence scores and use the seed labels’ dataset class IDs. Visual prompts are not supported for batch annotation.
- 🔎 More control over similar-image searches: A new
include_source_datasetoption lets users include the source dataset and datasets containing copies of the query image. Copies of the query image itself remain excluded. - ℹ️ Clearer partial-result guidance: The API now explains that reaching the detection cap may mean more objects exist, and that truncated generative output can omit objects or classes.
- 📦 Synchronized release: The OpenAPI contract and Python SDK were updated to version 0.1.90, with CI validating the regenerated SDK.
🎯 Purpose & Impact
- Visual prompts let users guide predictions with existing labels on an image, supporting more targeted detection, segmentation, and mask-derived oriented boxes.
- The new search option makes it easier to find related images within a chosen dataset scope while continuing to filter out duplicate query images.
- The synchronized contract and SDK give Python developers an up-to-date, validated interface for these API features.
What's Changed
- Update Platform API contract by @UltralyticsAssistant in #119
Full Changelog: v0.1.89...v0.1.90