GitHub - DukeRobotics/cv-training: Scripts to train computer vision models.

Skip to content

Navigation Menu

Sign in

Appearance settings

cv-training

Scripts to train computer vision models.

Dependencies

Ensure that you have a CUDA-enabled GPU. Only Ubuntu 22.04 LTS is officially supported.

To create the cv-train conda environment, run:

conda create --name cv-train python=3.11

Select the cv-train kernel when executing any Jupyter Notebooks in this repository.

Training

Use the yolov7_tiny.ipynb Jupyter Notebook to train a YOLOv7-tiny model.

CV Workflow

For a complete guide on our CV workflow, visit Computer Vision Workflow on our wiki.