public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-01 03:59:20 +02:00
Code Issues Packages Projects Releases Wiki Activity

whisper : add context param to disable gpu (#1293)

Browse Source 
* whisper : check state->ctx_metal not null

* whisper : add whisper_context_params { use_gpu }

* whisper : new API with params & deprecate old API

* examples : use no-gpu param && whisper_init_from_file_with_params

* whisper.objc : enable metal & disable on simulator

* whisper.swiftui, metal : enable metal & support load default.metallib

* whisper.android : use new API

* bindings : use new API

* addon.node : fix build & test

* bindings : updata java binding

* bindings : add missing whisper_context_default_params_by_ref WHISPER_API for java

* metal : use SWIFTPM_MODULE_BUNDLE for GGML_SWIFT and reuse library load

* metal : move bundle var into block

* metal : use SWIFT_PACKAGE instead of GGML_SWIFT

* style : minor updates

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Jhen-Jie Hong
2023-11-06 11:04:24 +02:00
committed by GitHub
co-authored by Georgi Gerganov
parent 39cfad0dee
commit 0463028bc2
29 changed files with 439 additions and 188 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/bench/bench.cpp
+11 -4
View File
@@ -11,6 +11,8 @@ struct whisper_params {
int32_t what = 0; // what to benchmark: 0 - whisper ecoder, 1 - memcpy, 2 - ggml_mul_mat
std::string model = "models/ggml-base.en.bin";
bool use_gpu = true;
};
void whisper_print_usage(int argc, char ** argv, const whisper_params & params);
@@ -23,9 +25,10 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) {
whisper_print_usage(argc, argv, params);
exit(0);
}
else if (arg == "-t" || arg == "--threads") { params.n_threads = std::stoi(argv[++i]); }
else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; }
else if (arg == "-w" || arg == "--what") { params.what = atoi(argv[++i]); }
else if (arg == "-t" || arg == "--threads") { params.n_threads = std::stoi(argv[++i]); }
else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; }
else if (arg == "-w" || arg == "--what") { params.what = atoi(argv[++i]); }
else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; }
else {
fprintf(stderr, "error: unknown argument: %s\n", arg.c_str());
whisper_print_usage(argc, argv, params);
@@ -45,6 +48,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para
fprintf(stderr, " -t N, --threads N [%-7d] number of threads to use during computation\n", params.n_threads);
fprintf(stderr, " -m FNAME, --model FNAME [%-7s] model path\n", params.model.c_str());
fprintf(stderr, " -w N, --what N [%-7d] what to benchmark:\n", params.what);
fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true");
fprintf(stderr, " %-7s 0 - whisper\n", "");
fprintf(stderr, " %-7s 1 - memcpy\n", "");
fprintf(stderr, " %-7s 2 - ggml_mul_mat\n", "");
@@ -54,7 +58,10 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para
int whisper_bench_full(const whisper_params & params) {
// whisper init
struct whisper_context * ctx = whisper_init_from_file(params.model.c_str());
struct whisper_context_params cparams;
cparams.use_gpu = params.use_gpu;
struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams);
{
fprintf(stderr, "\n");