I find Arm Neon has some usage in Pytorch. But I find the usage is usually for vec256 256 bits vector. But Arm Neon with android supports 128 bits vector at most. Does this condition will be implemented in the future?
I find Arm Neon has some usage in Pytorch. But I find the usage is usually for vec256 256 bits vector. But Arm Neon with android supports 128 bits vector at most. Does this condition will be implemented in the future?