Distribute and run LLMs with a single file.
C++ cross-platformggufllama-cpplocal-ailocal-inferencelocal-llmopen-source-aisingle-file-executablespeech-to-text
Make --nocompile flag work as intended
Justine Tunney474b44fe7c25
diff --git a/llamafile/cuda.c b/llamafile/cuda.cindex 4a08ee8a..c265df8c 100644--- a/llamafile/cuda.c+++ b/llamafile/cuda.c@@ -701,7 +701,9 @@ static bool ImportCudaImpl(void) {strlcat(dso, "ggml-rocm.", PATH_MAX);strlcat(dso, GetDsoExtension(), PATH_MAX);if (FLAG_nocompile) {- if (LinkCudaDso(dso, library_path)) {+ if ((FileExists(dso) ||+ ExtractCudaDso(dso, "ggml-rocm")) &&+ LinkCudaDso(dso, library_path)) {return true;} else {goto TryNvidia;@@ -776,7 +778,9 @@ TryNvidia:strlcat(dso, "ggml-cuda.", PATH_MAX);strlcat(dso, GetDsoExtension(), PATH_MAX);if (FLAG_nocompile) {- return LinkCudaDso(dso, library_path);+ return ((FileExists(dso) ||+ ExtractCudaDso(dso, "ggml-cuda")) &&+ LinkCudaDso(dso, library_path));}// Check if DSO is already compiled.