ultralytics / ultralytics

Ultralytics YOLO11 πŸš€
https://docs.ultralytics.com
GNU Affero General Public License v3.0
36.33k stars 7k forks source link

task=track #18955

Open JohnAsath opened 1 week ago

JohnAsath commented 1 week ago

Search before asking

Question

  1. !yolo task=detect mode=train model=yolov8m.pt data=/content/drive/MyDrive/Tuck-in_Tuck-out/data.yaml epochs=300 imgsz=1280 batch=8 patience=10 project=/content/drive/MyDrive/Tuck-in_Tuck-out/results name=test

instead of task = detect, task=track what will happen?

  1. How should the data be formatted for tracking, and what should the TXT format be?

Additional

No response

UltralyticsAssistant commented 1 week ago

πŸ‘‹ Hello @JohnAsath, thank you for your interest in Ultralytics πŸš€! We recommend exploring our Docs where you can find detailed guidance on various tasks, including detection and tracking, as well as Python and CLI usage examples. Many of the most common questions are addressed there!

If this is a πŸ› Bug Report, please provide a minimum reproducible example to help us debug it.

If this is a custom training or tracking ❓ Question, please include as much information as possible about your setup, such as your training command, how your data is structured, and any logs or errors you are encountering. For tracking-specific tasks, ensure your data adheres to the tracking dataset format.

For assistance with tracking-related questions specifically:

  1. Switching from task=detect to task=track changes the behavior to track object movements across frames. See the Tracking Guide for more information.
  2. Ensure your data is formatted correctly for tracking. Refer to the required TXT annotation structure and the MOT-compliant dataset details.

Join the Ultralytics community where it suits you best for more help and discussions. For real-time support, check out Discord 🎧. Prefer more in-depth conversations? Visit our Discourse. Or share knowledge and questions with others on our Subreddit 🌟.

Upgrade

Upgrade to the latest ultralytics package to ensure you’re working with the most up-to-date features and fixes. Install it in a Python>=3.8 environment with:

pip install -U ultralytics

Environments

YOLO, including tracking tasks, runs seamlessly in the following environments with dependencies (e.g., CUDA/CUDNN, Python, PyTorch) pre-installed:

Status

Ultralytics CI

If this badge is green, all Ultralytics CI tests are currently passing. CI tests ensure proper operation of YOLO for all Modes and Tasks across macOS, Windows, and Ubuntu.

This is an automated response to help guide you πŸ‘Ύ. An Ultralytics engineer will review your issue and provide personalized assistance as soon as possible. Stay tuned! 😊

Y-T-G commented 1 week ago
  1. Nothing probably. Or will default to detect based on the model being trained.
  2. There's no support for training using tracking datasets