diff --git a/README.md b/README.md index 1d4e9d417..72b5a2c88 100644 --- a/README.md +++ b/README.md @@ -1117,7 +1117,9 @@ docker run --gpus all -v /path/to/models:/models local/llama.cpp:server-cuda -m - Clean-up any trailing whitespaces, use 4 spaces for indentation, brackets on the same line, `void * ptr`, `int & a` - See [good first issues](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) for tasks suitable for first contributions - Tensors store data in row-major order. We refer to dimension 0 as columns, 1 as rows, 2 as matrices -- Matrix multiplication is unconventional: [`z = ggml_mul_mat(ctx, x, y)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means `zT = x @ yT` +- Matrix multiplication is unconventional: [`C = ggml_mul_mat(ctx, A, B)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means $C^T = A B^T \Leftrightarrow C = B A^T.$ + +![matmul](media/matmul.png) ### Docs diff --git a/media/matmul.png b/media/matmul.png new file mode 100644 index 000000000..786a20492 Binary files /dev/null and b/media/matmul.png differ diff --git a/media/matmul.svg b/media/matmul.svg new file mode 100644 index 000000000..1d6cb4bb7 --- /dev/null +++ b/media/matmul.svg @@ -0,0 +1,1238 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + ARow-major + BTColumn-major + CT=ABTColumn-major + + ne00 + + ne01 + + ne1 + + ne0 + + ne10 + + ne11 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + BRow-major + ATColumn-major + C=BATRow-major + + ne10 + + ne11 + + ne0 + + ne1 + + ne00 + + ne01 + + +