Distribute and run LLMs with a single file.
C++ cross-platformggufllama-cpplocal-ailocal-inferencelocal-llmopen-source-aisingle-file-executablespeech-to-text
Redraw tinyBLAS ANSI art logo
Justine Tunneyd10eb328b82c
diff --git a/llamafile/tinyblas_cpu_mixmul.inc b/llamafile/tinyblas_cpu_mixmul.incindex 9e42d874..f0a73e7f 100644--- a/llamafile/tinyblas_cpu_mixmul.inc+++ b/llamafile/tinyblas_cpu_mixmul.inc@@ -18,13 +18,15 @@#include "tinyblas_cpu.h"//-// _ _ ___ _ _ ___-// | |_(_)_ _ _ _| _ ) | /_\ / __|-// | _| | ' \ || | _ \ |__ / _ \\__ \.-// \__|_|_||_\_, |___/____/_/ \_\___/-// |__///-// MIXTURE OF EXPERTS TENSOR MULTIPLICATION+// ██████╗ ██╗ █████╗ ██████╗+// ██████╗██╗██╗ ██╗██═██╗██╔══██╗██║ ██╔══██╗██╔═══╝+// ╚═██╔═╝██║███▄██║██ ██║██████╔╝██║ ███████║██████╗+// ██║ ██║██▀███║╚███╔╝██╔══██╗██║ ██╔══██║╔═══██║+// ██║ ██║██║ ██║ ███║ ██████╔╝████╗██║ ██║██████║+// ╚═╝ ╚═╝╚═╝ ╚═╝ ╚══╝ ╚═════╝ ╚═══╝╚═╝ ╚═╝╚═════╝+//+// MIXTURE OF EXPERTS TENSOR MULTIPLICATION////// SHAPESdiff --git a/llamafile/tinyblas_cpu_sgemm.inc b/llamafile/tinyblas_cpu_sgemm.incindex 1adda4f8..9cbb6d1b 100644--- a/llamafile/tinyblas_cpu_sgemm.inc+++ b/llamafile/tinyblas_cpu_sgemm.inc@@ -18,13 +18,15 @@#include "tinyblas_cpu.h"//-// _ _ ___ _ _ ___-// | |_(_)_ _ _ _| _ ) | /_\ / __|-// | _| | ' \ || | _ \ |__ / _ \\__ \.-// \__|_|_||_\_, |___/____/_/ \_\___/-// |__///-// BASIC LINEAR ALGEBRA SUBPROGRAMS+// ██████╗ ██╗ █████╗ ██████╗+// ██████╗██╗██╗ ██╗██═██╗██╔══██╗██║ ██╔══██╗██╔═══╝+// ╚═██╔═╝██║███▄██║██ ██║██████╔╝██║ ███████║██████╗+// ██║ ██║██▀███║╚███╔╝██╔══██╗██║ ██╔══██║╔═══██║+// ██║ ██║██║ ██║ ███║ ██████╔╝████╗██║ ██║██████║+// ╚═╝ ╚═╝╚═╝ ╚═╝ ╚══╝ ╚═════╝ ╚═══╝╚═╝ ╚═╝╚═════╝+//+// BASIC LINEAR ALGEBRA SUBPROGRAMS////// This file implements multithreaded CPU matrix multiplication for the@@ -321,7 +323,7 @@ bool llamafile_sgemm(long m, long n, long k, const void *A, long lda, const void}}if ((Btype == GGML_TYPE_Q8_0 || Btype == GGML_TYPE_Q8_1) && Ctype == GGML_TYPE_F32) {- assert (QK8_0 == QK8_1 == QK4_0 == QK4_1 == QK5_0 == QK5_1 == 32);+ assert(QK8_0 == QK8_1 == QK4_0 == QK4_1 == QK5_0 == QK5_1 == 32);if (iqk_mul_mat(m, n, k * QK8_0, Atype, A, B, (float *)C, ldc, ith, nth)) {return true;}