Distribute and run LLMs with a single file.
C++ cross-platformggufllama-cpplocal-ailocal-inferencelocal-llmopen-source-aisingle-file-executablespeech-to-text
Disable GPU in whisperfile by default
Justine Tunney2841793a94aa
diff --git a/whisper.cpp/main.cpp b/whisper.cpp/main.cppindex 8ff8bbc3..4cfd75bb 100644--- a/whisper.cpp/main.cpp+++ b/whisper.cpp/main.cpp@@ -939,6 +939,9 @@ static void cb_log_disable(enum ggml_log_level , const char * , void * ) { }int main(int argc, char ** argv) {whisper_params params;+ // [jart] disable GPU by default (pass `--gpu auto` to enable it)+ FLAG_gpu = LLAMAFILE_GPU_DISABLE;+// If the only argument starts with "@", read arguments line-by-line// from the given file.std::vector<std::string> vec_args;diff --git a/whisper.cpp/server.cpp b/whisper.cpp/server.cppindex fbd1246e..6c5feef4 100644--- a/whisper.cpp/server.cpp+++ b/whisper.cpp/server.cpp@@ -505,6 +505,9 @@ int main(int argc, char ** argv) {std::mutex whisper_mutex;+ // [jart] disable GPU by default (pass `--gpu auto` to enable it)+ FLAG_gpu = LLAMAFILE_GPU_DISABLE;+LoadZipArgs(&argc, &argv);if (whisper_params_parse(argc, argv, params, sparams) == false) {