shirsh10mall/Object-Detection-for-Traffic-Vehicle-Detection — reverse-engineered prompt
Reverse engineered prompt
Build me a simple traffic vehicle detection project that I can run locally and demo easily. I want a small web app where someone can upload a road or CCTV image, click a button, and see the same image back with the detected vehicles outlined and labeled. Please use the trained model and assets that are already in this repo if possible, and keep the vehicle classes aligned with the dataset here, like two wheelers, LMV, HMV, and auto rickshaws.
Also make sure the training and deployment notebooks still make sense, so someone can understand how the data was prepared, how the model was trained, and how the web app uses the saved weights. If anything is missing or broken, fix it so the whole project feels complete and runnable from this repo. I would also like the app UI to be clean and beginner friendly, similar to the sample screenshots. If you need details, check current docs online and fill in the gaps.
Want more depth? Deep Reverse