Plan and build a computer vision project with a prompt that covers the whole path from dataset to deployment. It is aimed at developers who describe a CV task and want an architecture recommendation with code fragments.
What you get
- An architecture recommendation matched to the CV task you describe
- Dataset preparation guidance, including where to start when your dataset is small
- Augmentation suggestions for your data
- OpenCV and PyTorch code fragments, such as a setup with a pretrained backbone
- Guidance on exporting and deploying the trained model, not only training it
- A plain-text file, ai-computer-vision-assistant-skill.txt
Who it's for
Developers working on image or video tasks who are comfortable with code and want a second opinion on structure and next steps. The prompt is set at an advanced level: it assumes you can run PyTorch and OpenCV yourself and want specific recommendations tied to your dataset and goal rather than a general tutorial on computer vision.
Works with Claude, ChatGPT and other AI chats.