Crates.io | whisper-rs |
lib.rs | whisper-rs |
version | 0.15.1 |
created_at | 2022-10-10 03:00:57.021201+00 |
updated_at | 2025-09-10 22:37:32.734097+00 |
description | Rust bindings for whisper.cpp |
homepage | |
repository | https://codeberg.org/tazz4843/whisper-rs |
max_upload_size | |
id | 684326 |
size | 169,337 |
Rust bindings to whisper.cpp
git clone --recursive https://codeberg.org/tazz4843/whisper-rs.git
cd whisper-rs
cargo run --example basic_use
cargo run --example audio_transcription
use whisper_rs::{WhisperContext, WhisperContextParameters, FullParams, SamplingStrategy};
fn main() {
let path_to_model = std::env::args().nth(1).unwrap();
// load a context and model
let ctx = WhisperContext::new_with_params(
path_to_model,
WhisperContextParameters::default()
).expect("failed to load model");
// create a params object
let params = FullParams::new(SamplingStrategy::BeamSearch {
beam_size: 5,
patience: -1.0,
});
// assume we have a buffer of audio data
// here we'll make a fake one, floating point samples, 32 bit, 16KHz, mono
let audio_data = vec![0_f32; 16000 * 2];
// now we can run the model
let mut state = ctx.create_state().expect("failed to create state");
state
.full(params, &audio_data[..])
.expect("failed to run model");
// fetch the results
for segment in state.as_iter() {
println!(
"[{} - {}]: {}",
// note start and end timestamps are in centiseconds
// (10s of milliseconds)
segment.start_timestamp(),
segment.end_timestamp(),
// the Display impl for WhisperSegment will replace invalid UTF-8 with the Unicode replacement character
segment
);
}
}
See examples/basic_use.rs for more details.
Lower level bindings are exposed if needed, but the above should be enough for most use cases. See the docs: https://docs.rs/whisper-rs/ for more details.
All disabled by default unless otherwise specified.
raw-api
: expose whisper-rs-sys without having to pull it in as a dependency.
NOTE: enabling this no longer guarantees semver compliance,
as whisper-rs-sys may be upgraded to a breaking version in a patch release of whisper-rs.cuda
: enable CUDA support. Implicitly enables hidden GPU flag at runtime.hipblas
: enable ROCm/hipBLAS support. Only available on linux. Implicitly enables hidden GPU flag at runtime.openblas
: enable OpenBLAS support.metal
: enable Metal support. Implicitly enables hidden GPU flag at runtime.vulkan
: enable Vulkan support. Implicitly enables hidden GPU flag at runtime.log_backend
: allows hooking into whisper.cpp's log output and sending it to the log
backend. Requires callingtracing_backend
: allows hooking into whisper.cpp's log output and sending it to the tracing
backend.See BUILDING.md for instructions for building whisper-rs on Windows and OSX M1. Linux builds should just work out of the box.
WHISPER_DONT_GENERATE_BINDINGS
environment variable.
This skips attempting to build the bindings whatsoever and copies the existing ones. They may be out of date,
but it's better than nothing.
WHISPER_DONT_GENERATE_BINDINGS=1 cargo build
tl;dr: code is in the public domain
the PR template is derived from GoToSocial's PR template