Benchmark of Apple's MLX operations on mlx gpu, cpu, torch mps and cuda. - View it on GitHub