Skip to content

Remove model_name param from Whisper-Metal #15798

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Nov 15, 2025
+4 −19
Merged

Remove model_name param from Whisper-Metal #15798

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ed8786f
feat: no need to specify decoder_start_token_id
seyeong-han Nov 12, 2025
16685b5
Merge branch 'main' into remove-modelname-whisper-metal
seyeong-han Nov 14, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 4 additions & 19 deletions examples/models/whisper/main.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,10 +39,6 @@ DEFINE_string(
audio_path,
"",
"Path to input audio file. Accepts .wav or raw float .bin.");
DEFINE_string(
model_name,
"base",
"Whisper model name (base, small, medium, large, large-v2, large-v3, large-v3-turbo).");
DEFINE_double(
temperature,
0.0,
Expand Down Expand Up @@ -114,21 +110,10 @@ int main(int argc, char** argv) {
config.max_new_tokens = FLAGS_max_new_tokens;
config.temperature = static_cast<float>(FLAGS_temperature);

// Set decoder_start_token_id based on model version
if (FLAGS_model_name == "large-v2" || FLAGS_model_name == "large-v3" ||
FLAGS_model_name == "large-v3-turbo") {
config.decoder_start_token_id = 50258;
ET_LOG(
Info,
"Using decoder_start_token_id=50258 for model: %s",
FLAGS_model_name.c_str());
} else {
config.decoder_start_token_id = 50257;
ET_LOG(
Info,
"Using decoder_start_token_id=50257 for model: %s",
FLAGS_model_name.c_str());
}
// All Whisper models from HuggingFace now use the v3 tokenizer format
// where token 50257 = <|endoftext|> and token 50258 = <|startoftranscript|>
config.decoder_start_token_id = 50258;
ET_LOG(Info, "Using decoder_start_token_id=50258");

auto result =
runner.transcribe(features, config, [&](const std::string& piece) {
Expand Down
Loading