盲盒3D推理 all-round-chibi

Update log：
V3: Input a single image with a clear subject，Can generate images of the subject rotated to different orientations（Tic-tac-toe grid*1 sheet）
V2: Input a single image with a clear subject（Workflow automatically removes background），No other parameters required，Can generate images of the subject rotated to different orientations（*6 photos）
V1: Input a single image with a simple background and a clear subject，Pitch angle/Azimuth angle（0.1±180.0），Can generate an image of the subject rotated to a specified orientation（1 sheet）
Update log:
V3: Input an image of a single clear subject, and you can get the image of the subject after rotating to different positions (nine-grid image * 1 picture)
V2: Input an image of a single clear subject (the workflow automatically deducts the background), and you can get the image of the subject after rotating to different positions without other parameters (*6 pictures)
V1: Input a simple background image of a single clear subject, and the elevation/azimuth angle (0.1±180.0), and you can get the image of the subject after rotating to a specified position (1 picture)