TorchVision由C++(CUDA)和Python3实现**,依赖Torch、PNG、JPEG,还依赖PIL(Pillow, ... data/image/1.jpg") print("tensor shape:", tensor.shape) ...
確定! 回上一頁