# New Release: Ultralytics v8.3.67

**URL:** <https://community.ultralytics.com/t/new-release-ultralytics-v8-3-67/714>\
**Category:** Discussion\
**Tags:** ultralytics-official, releases, announcements\
**Created:** [January 24, 2025, 1:19pm UTC](https://community.ultralytics.com/t/new-release-ultralytics-v8-3-67/714 "2025-01-24T13:19:45Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![glenn-jocher](https://sea1.discourse-cdn.com/flex001/user_avatar/community.ultralytics.com/glenn-jocher/32/83_2.png) [@glenn-jocher](https://community.ultralytics.com/u/glenn-jocher)\
**Post date:** [January 24, 2025, 1:19pm UTC](https://community.ultralytics.com/t/new-release-ultralytics-v8-3-67/714/1 "2025-01-24T13:19:45Z")

</div>

# 🚀 Ultralytics v8.3.67 Release Announcement 🎉

Hello YOLO Community!

We’re excited to announce the release of **Ultralytics v8.3.67** , packed with incredible enhancements and updates that make developing and deploying YOLO models smoother than ever before.

## 🌟 Summary

v8.3.67 brings **Non-Maximum Suppression (NMS)** export capabilities to all YOLO tasks, simplifying deployment pipelines and improving portability and performance for real-time applications across a wide range of frameworks.

* * *

## 📢 New Features

### 1️⃣ NMS Support for YOLO Model Exports

- Export models with integrated **Non-Maximum Suppression (NMS)** for easier deployment 🔗.
- Supported formats: **ONNX** , **TensorRT** , **TFLite** , **TFJS** , **SavedModel** , **OpenVINO** , and **TorchScript** 🧩.
- Applicable to all tasks: **Detection** , **Segmentation** , **Pose Estimation** , and **Oriented Bounding Box (OBB)**.
- New export options, including `nms=True`, to enable task-specific NMS configurations.

### 2️⃣ Enhanced NMS Functionality in Models

- **NMSModel** wrapper for seamless embedding of NMS in exported models.
- Support for advanced configurations like **agnostic NMS** and **rotated NMS** for OBB tasks 🔄.
- **Streamlined APIs** to simplify the integration of NMS into your deployment pipelines.

* * *

## 🎯 Purpose & Impact

### Purpose

- Save time and effort by automating post-processing pipelines 🕒.
- Simplify real-time deployment workflows for platforms like TensorFlow, ONNX, and Edge TPU 🚀.

### Impact

- Improved **portability** for cross-platform deployments 📦.
- Enhanced **robustness** by unifying pre/post-processing pipelines.
- Reduced complexity and potential **errors** in real-world applications.

Take your YOLO deployments to the next level with v8.3.67 – faster, smarter, and cleaner! 💪

* * *

## 🔧 Improvements

- Updated the **HUB Shared Inference API limits** for smoother usability ([#18850](https://github.com/ultralytics/ultralytics/pull/18850) by [@sergiuwaxmann](https://github.com/sergiuwaxmann)).
- Introduced a new `YOLO_TQDM_RICH` environment variable for improved progress bar handling ([#18854](https://github.com/ultralytics/ultralytics/pull/18854) by [@glenn-jocher](https://github.com/glenn-jocher)).
- Complete overhaul of the export pipeline to include **task-specific NMS** ([#18484](https://github.com/ultralytics/ultralytics/pull/18484) by [@Y-T-G](https://github.com/Y-T-G)).

* * *

## 📝 Full Changelog and Release

Check out the full changelog here: [v8.3.67 Changelog](https://github.com/ultralytics/ultralytics/compare/v8.3.66...v8.3.67)  
View and download the release: [v8.3.67 Release](https://github.com/ultralytics/ultralytics/releases/tag/v8.3.67)

* * *

## 💬 Feedback and Suggestions

We’d love to hear from you! Try out these new features and let us know what you think. Your feedback helps us shape the future of Ultralytics YOLO. Drop by our [Discussions](https://github.com/ultralytics/ultralytics/discussions) to share your thoughts or ask questions!

* * *

Thank you for being an amazing community and pushing the boundaries of computer vision! This update is a step forward in making YOLO models more powerful and accessible across all platforms and tasks. We couldn’t do this without your support and collaboration.

Happy YOLO-ing! 🚀  
— The Ultralytics Team