When writing the optimization program of cubic curving, the Jacobian matrix needs to be recalculated repeatedly due to the need to test the possible results of the constraints of different conditions, which is very time-consuming and prone to errors. So we need an automatic, fast and accurate derivation method. This project refers to the back-propagation calculation method of pytorch and is equipped with Jacobian matrix index management to realize automatic Jacobian matrix solving
If you encounter any problems during use, please contact me