2 ms·Running PyTorch Models on Apple Silicon GPUs with the ExecuTorch MLX Delegate2 points by 0bytematt 5mo ago