KieDani/UpliftingTableTennis ? reverse-engineered prompt
Reverse engineered prompt
Build me a simple way to run this table tennis project on my own videos from start to finish.
I want to point it at a match video, detect the ball and table, estimate the 3D trajectory and spin, and then save the results in an easy to view format like annotated video and a small output file with the tracked data. If there’s already an existing inference flow here, please wire it up into one clean command or a small interface so I don’t have to figure out all the separate scripts.
Please make it practical for someone who just wants to use the pretrained models, and include clear setup steps or defaults for the required paths. If anything depends on current library versions or docs, look up the latest guidance online if you need to.
Are you gonna build this?
make sure you review the code using coderabbit