ru/models/sam-2/

[giscus[bot]]

ru/models/sam-2/

Откройте для себя SAM 2, следующее поколение модели Segment Anything Model от Meta, поддерживающее сегментацию в реальном времени на изображениях и видео с самой современной производительностью. Узнайте о ее ключевых особенностях, наборах данных и способах использования.

https://docs.ultralytics.com/ru/models/sam-2/

[Alex-mtnkv]

Hello, Ultralytics team!
Is there a plan to export to ONNX and TensorRT?

[glenn-jocher]

@Alex-mtnkv yes, exporting to ONNX and TensorRT is supported in Ultralytics. You can use the ultralytics export command or the Python API to convert models into these formats. Refer to the Export Documentation for detailed instructions.

[Alex-mtnkv]

For SAM2?

[glenn-jocher]

Yes, SAM2, as well as the newer SAM2.1, is supported. You can find detailed information on its features, usage, and export options in the SAM 2 Documentation.

[UltralyticsAssistant]

👋 Hello, thank you for sharing your interest in SAM 2 🚀! The Segment Anything Model (SAM 2) from Meta indeed unlocks powerful capabilities in segmentation for both images and videos. We recommend taking a closer look at the information provided in the Ultralytics Docs for a deep dive into its features, datasets, and applications.

If this is a 🐛 Bug Report, please provide a minimum reproducible example so that we can better assist you in debugging the issue.

If this is a custom training ❓ Question or you'd like to explore SAM 2 with specific use cases, please share more details about your setup, dataset, desired outcomes, or any logs so we can guide you effectively. Don't forget to check out our Tips for Best Training Results for useful insights.

Join the vibrant Ultralytics community to discuss with other AI enthusiasts. Chat with us in real time on Discord 🎧, dive into in-depth conversations on Discourse, or explore our Subreddit to learn and share more.

Upgrade

To ensure compatibility and take full advantage of our latest features and improvements, upgrade to the newest ultralytics release in a Python>=3.8 environment with PyTorch>=1.8:

pip install -U ultralytics

Environments

Try YOLO and SAM 2 in various verified environments equipped with pre-installed dependencies like CUDA, Python, and PyTorch:

• Notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

The Ultralytics CI badge above reflects the latest test results. A green badge indicates passing tests for all YOLO Modes and Tasks on supported platforms.

📢 This is an automated response to get you started. An Ultralytics engineer will assist you further soon!