Rotated Mnist

Rotated MNIST is a variant of the MNIST dataset, where the original handwritten digit images are rotated to increase data diversity. The task aims to evaluate and improve the performance of computer vision models in handling rotation invariance issues, thereby enhancing the model's generalization and robustness. In practical applications, Rotated MNIST helps to increase the accuracy and stability of image recognition systems at different angles and orientations.