Whisper-rs

https://docs.rs/whisper-rs/latest/whisper_rs/
Admin
Whisper-rs is a Rust wrapper for the `whisper.cpp` library, providing a full transcription pipeline,...
Tokens:34,154
Snippets:424
Trust Score:9.7
Update:1 month ago
Show doc for...
Tokens:
Raw
### Example Usage of convert_stereo_to_mono_audio

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.convert_stereo_to_mono_audio.html?search=

Demonstrates how to initialize buffers and invoke the conversion function.

```rust
let samples = [0.0f32; 1024];
let mut mono_samples = [0.0f32; 512];
convert_stereo_to_mono_audio(&samples, &mut mono_samples).expect("should be no half samples missing");
```

--------------------------------

### Get Audio Context Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the audio context size.

```rust
pub fn n_audio_ctx(&self) -> c_int
```

```c
int whisper_n_audio_ctx     (struct whisper_context * ctx);
```

--------------------------------

### Usage Example for convert_integer_to_float_audio

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.convert_integer_to_float_audio.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Demonstrates how to prepare buffers and invoke the conversion function. Note that the function will panic if the input and output lengths do not match.

```rust
let samples = [0i16; 1024];
let mut output = vec![0.0f32; samples.len()];
convert_integer_to_float_audio(&samples, &mut output).expect("input and output lengths should be equal");
```

--------------------------------

### Get Text Context Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the text context size.

```rust
pub fn n_text_ctx(&self) -> c_int
```

```c
int whisper_n_text_ctx     (struct whisper_context * ctx);
```

--------------------------------

### Get Vocabulary Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the size of the vocabulary.

```rust
pub fn n_vocab(&self) -> c_int
```

```c
int whisper_n_vocab        (struct whisper_context * ctx)
```

--------------------------------

### Get Segment Start Timestamp

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the start time of the segment in centiseconds (10s of milliseconds).

```rust
pub fn start_timestamp(&self) -> i64
```

--------------------------------

### install_logging_hooks function signature

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.install_logging_hooks.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

The function signature for installing logging hooks in whisper-rs.

```rust
pub fn install_logging_hooks()
```

--------------------------------

### Install Logging Hooks

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.install_logging_hooks.html

Redirects whisper.cpp and GGML logs to the Rust logging system. Enable `log_backend` or `tracing_backend` features to direct logs to `log` or `tracing` respectively. Without these, logs are effectively disabled. Safe to call multiple times.

```rust
pub fn install_logging_hooks()
```

--------------------------------

### Get Model Type

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the model's type.

```rust
pub fn model_type(&self) -> c_int
```

```c
int whisper_model_type         (struct whisper_context * ctx);
```

--------------------------------

### Get Whisper Version

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.get_whisper_version.html

Retrieves the static string slice representing the current whisper.cpp version. No setup or imports are required beyond the library itself.

```rust
pub fn get_whisper_version() -> &'static str
```

--------------------------------

### install_logging_hooks

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.install_logging_hooks.html

Redirects all whisper.cpp and GGML logs to logging hooks installed by whisper-rs. This function is safe to call multiple times, but only has an effect the first time it is invoked.

```APIDOC
## install_logging_hooks

### Description
Redirects all whisper.cpp and GGML logs to logging hooks installed by whisper-rs. This integration allows logs to be processed by `log` or `tracing` frameworks if the corresponding features are enabled. If neither feature is enabled, logs will effectively be disabled.

Note that whisper.cpp and GGML do not consistently adhere to Rust logging conventions. Control log output using your logging crate's configuration. It is recommended to configure by module path and use `whisper_rs::ggml_logging_hook` and/or `whisper_rs::whisper_logging_hook` to ensure future `whisper-rs` logs are not inadvertently ignored.

This function is safe to call multiple times; it only affects the system on its first invocation. Subsequent calls have no additional effect.

### Function Signature
```rust
pub fn install_logging_hooks()
```

### Usage
This function does not take any parameters and does not return any value. Its effect is to modify the global logging behavior for the whisper.cpp and GGML libraries.
```

--------------------------------

### Get Model Vocabulary Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the model's vocabulary size.

```rust
pub fn model_n_vocab(&self) -> c_int
```

```c
int whisper_model_n_vocab      (struct whisper_context * ctx);
```

--------------------------------

### Get Audio Context Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the maximum number of audio samples (or frames) the model can process at once. This defines the window size for audio input.

```rust
pub fn n_audio_ctx(&self) -> c_int
```

--------------------------------

### Get Readable Model Type

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=std%3A%3Avec

Retrieves the model type as a byte slice.

```rust
pub fn model_type_readable_bytes(&self) -> Result<&[u8], WhisperError>
```

```cpp
const char * whisper_model_type_readable(struct whisper_context * ctx);
```

--------------------------------

### Get Model Audio Context Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Returns the audio context size as defined by the model architecture. This is the number of audio frames the model's encoder can handle.

```rust
pub fn model_n_audio_ctx(&self) -> c_int
```

--------------------------------

### Get Model File Type

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves an integer code representing the file type or precision of the loaded model (e.g., f32, f16).

```rust
pub fn model_ftype(&self) -> c_int
```

--------------------------------

### Get Vocabulary Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the number of tokens in the model's vocabulary. This value is constant for a given model.

```rust
pub fn n_vocab(&self) -> c_int
```

--------------------------------

### Get Model Text Context Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Returns the text context size as defined by the model architecture. This is the maximum number of tokens the text decoder can consider.

```rust
pub fn model_n_text_ctx(&self) -> c_int
```

--------------------------------

### Convert Stereo to Mono Audio

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.convert_stereo_to_mono_audio.html

Converts 32-bit floating point stereo PCM audio to mono PCM audio. Ensure the input array length is even and the output array has half the capacity of the input. This example demonstrates a successful conversion with pre-allocated buffers.

```rust
let samples = [0.0f32; 1024];
let mut mono_samples = [0.0f32; 512];
convert_stereo_to_mono_audio(&samples, &mut mono_samples).expect("should be no half samples missing");
```

--------------------------------

### Get Text Context Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Returns the maximum number of tokens the model can process in its text context. This indicates the model's memory for textual input.

```rust
pub fn n_text_ctx(&self) -> c_int
```

--------------------------------

### Get Model Vocabulary Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the vocabulary size specific to the model architecture. This might differ from the context's vocabulary size in some configurations.

```rust
pub fn model_n_vocab(&self) -> c_int
```

--------------------------------

### Get segment timestamps

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html?search=

Retrieves the start and end timestamps of the segment in centiseconds.

```rust
pub fn start_timestamp(&self) -> i64
```

```rust
pub fn end_timestamp(&self) -> i64
```

--------------------------------

### get_lang_str_full

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.get_lang_str_full.html

Gets the full string representation of a language ID. For example, an ID of 2 returns 'german'.

```APIDOC
## Function get_lang_str_full

### Description
Get the full string of the specified language name (e.g. 2 -> “german”).

### Signature
```rust
pub fn get_lang_str_full(id: i32) -> Option<&'static str>
```

### Parameters
* **id** (i32) - The ID of the language.

### Returns
- `Option<&'static str>`: The full string of the language, or `None` if the ID is not found.

### C++ Equivalent
`const char * whisper_lang_str_full(int id)`
```

--------------------------------

### Set Initial Prompt

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=

Sets the initial prompt text for the model. Panics if the provided string contains a null byte.

```rust
let mut params = FullParams::new(SamplingStrategy::Greedy { best_of: 5 });
params.set_initial_prompt("Hello, world!");
// ... further usage of params ...
```

--------------------------------

### Get Model Audio State Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the size of the internal audio state maintained by the model. This relates to the model's memory for audio processing.

```rust
pub fn model_n_audio_state(&self) -> c_int
```

--------------------------------

### full

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html

Runs the entire model on the provided PCM audio data using the specified decoding strategy.

```APIDOC
## pub fn full(&mut self, params: FullParams<'_, '_>, data: &[f32]) -> Result<(), WhisperError>

### Description
Runs the entire model: PCM to log mel spectrogram, encoder, and decoder to generate text. This is typically the primary function used by end users.

### Parameters
- **params** (FullParams) - Required - The decoding strategy parameters.
- **data** (&[f32]) - Required - Raw PCM audio data (32-bit float, 16kHz, 1 channel).

### Returns
- **Result<(), WhisperError>** - Returns Ok(()) on success, or a WhisperError on failure.
```

--------------------------------

### get_lang_str

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.get_lang_str.html

Gets the short string representation of a language ID. For example, an ID of 2 returns 'de'.

```APIDOC
## get_lang_str

### Description
Get the short string of the specified language id (e.g. 2 -> “de”).

### Signature
```rust
pub fn get_lang_str(id: i32) -> Option<&'static str>
```

### Parameters
#### Path Parameters
- **id** (i32) - Required - The language ID.

### Returns
- **Option<&'static str>** - The short string of the language, None if not found.

### C++ Equivalent
`const char * whisper_lang_str(int id)`
```

--------------------------------

### print_system_info()

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.print_system_info.html?search=

Prints system information and returns a static string reference.

```APIDOC
## Function: print_system_info

### Description
Prints system information to the console and returns a static string reference containing the information.

### Signature
`pub fn print_system_info() -> &'static str`

### Returns
- **&'static str** - A static string reference containing the system information.
```

--------------------------------

### FullParams Configuration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=u32+-%3E+bool

Methods for configuring callbacks, grammar, prompts, and VAD settings for the whisper model.

```APIDOC
## pub fn set_initial_prompt(&mut self, initial_prompt: &str)

### Description
Sets the initial prompt for the model, which serves as the starting point for the model's decoding process. Calling this multiple times will overwrite the previous prompt.

### Parameters
- **initial_prompt** (&str) - Required - The initial prompt text.

### Panics
Panics if the provided string contains a null byte.

---

## pub fn enable_vad(&mut self, vad: bool)

### Description
Enables or disables Voice Activity Detection (VAD).

### Parameters
- **vad** (bool) - Required - Whether to enable VAD.

### Panics
Panics if the VAD model path has not been set prior to enabling.

---

## pub fn set_vad_model_path(&mut self, vad_model_path: Option<&str>)

### Description
Sets the file path for the VAD model. Passing None disables VAD.

### Parameters
- **vad_model_path** (Option<&str>) - Required - Path to the VAD model file.

### Panics
Panics if the path contains a null byte.

---

## pub fn set_grammar(&mut self, grammar: Option<&[WhisperGrammarElement]>)

### Description
Enables an array of grammar elements to constrain the model output.

### Parameters
- **grammar** (Option<&[WhisperGrammarElement]>) - Required - The grammar elements to apply.

---

## pub fn set_grammar_penalty(&mut self, grammar_penalty: f32)

### Description
Sets the penalty applied to grammar violations.

### Parameters
- **grammar_penalty** (f32) - Required - The penalty value.
```

--------------------------------

### get_segment_start_timestamp

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadSegments.html?search=u32+-%3E+bool

Retrieves the start timestamp of a specific segment.

```APIDOC
## pub fn get_segment_start_timestamp(&self, idx: c_int) -> Option<f32>

### Description
Returns the start timestamp of the segment at the specified index in centiseconds (10s of milliseconds).

### Parameters
- **idx** (c_int) - The index of the segment.

### Returns
- **Option<f32>** - The start timestamp if the index is valid, otherwise None.
```

--------------------------------

### print_system_info

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.print_system_info.html?search=std%3A%3Avec

Prints system information to the console.

```APIDOC
## Function: print_system_info

### Description
Prints system information. This function is part of the whisper_rs crate.

### Signature
`pub fn print_system_info() -> &'static str`

### Returns
- **&'static str** - A static string reference containing system information.
```

--------------------------------

### Configuration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContextParameters.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Methods to configure the Whisper context parameters using a builder pattern.

```APIDOC
## Configuration Methods

### use_gpu(&mut self, use_gpu: bool) -> &mut Self
Configures whether to use the GPU if available.

### flash_attn(&mut self, flash_attn: bool) -> &mut Self
Enables or disables flash attention. Note: Cannot be used with DTW.

### gpu_device(&mut self, gpu_device: c_int) -> &mut Self
Sets the GPU device ID (default is 0).

### dtw_parameters(&mut self, dtw_parameters: DtwParameters<'a>) -> &mut Self
Sets the DTW token level timestamp parameters.
```

--------------------------------

### Create WhisperContext from File

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Creates a new WhisperContext by loading a model from a specified file path with custom parameters. Ensure the model file exists and parameters are correctly configured.

```rust
pub fn new_with_params<P>( path: P, parameters: WhisperContextParameters<'_>, ) -> Result<Self, WhisperError> where P: AsRef<Path>,
```

--------------------------------

### Get Mel Spectrogram Length

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the length of the mel spectrogram.

```rust
pub fn n_len(&self) -> c_int
```

```c
int whisper_n_len_from_state(struct whisper_context * ctx)
```

--------------------------------

### full

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=

Executes the full Whisper model pipeline on the provided PCM audio data using the specified decoding strategy.

```APIDOC
## pub fn full(&mut self, params: FullParams<'_, '_>, data: &[f32]) -> Result<(), WhisperError>

### Description
Runs the entire model: PCM to log mel spectrogram, encoder, and decoder to obtain text. This is the primary function for end users to process audio.

### Parameters
- **params** (FullParams) - Required - The decoding parameters.
- **data** (&[f32]) - Required - Raw PCM audio data, 32-bit floating point at 16 kHz, 1 channel.

### Returns
- **Result<(), WhisperError>** - Ok(()) on success, or a WhisperError on failure.
```

--------------------------------

### Get No Speech Probability

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the probability that the segment contains no speech.

```rust
pub fn no_speech_probability(&self) -> f32
```

--------------------------------

### print_system_info

Source: https://docs.rs/whisper-rs/latest/whisper_rs/index.html

Prints system information related to the whisper.cpp environment.

```APIDOC
## print_system_info

### Description
Prints system information related to the whisper.cpp environment.

### Method
(Not specified, likely a free function)

### Endpoint
(Not applicable, this is an SDK function)

### Parameters
(None)

### Request Example
(Not applicable)

### Response
(No return value specified)

#### Response Example
(Not applicable)
```

--------------------------------

### Get Model Audio Head Count

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Returns the number of attention heads in the audio processing layers of the model. This impacts the model's ability to focus on different parts of the audio.

```rust
pub fn model_n_audio_head(&self) -> c_int
```

--------------------------------

### Initialize WhisperContext

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=

Methods to create a new context from a file path or a byte buffer.

```rust
pub fn new_with_params<P>( path: P, parameters: WhisperContextParameters<'_>, ) -> Result<Self, WhisperError>
where P: AsRef<Path>,
```

```c
struct whisper_context * whisper_init_from_file_with_params_no_state(const char * path_model, struct whisper_context_params params);
```

```rust
pub fn new_from_buffer_with_params( buffer: &[u8], parameters: WhisperContextParameters<'_>, ) -> Result<Self, WhisperError>
```

```c
struct whisper_context * whisper_init_from_buffer_with_params_no_state(void * buffer, size_t buffer_size, struct whisper_context_params params);
```

--------------------------------

### WhisperVadContext::new

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadContext.html

Creates a new instance of WhisperVadContext.

```APIDOC
## pub fn new(model_path: &str, params: WhisperVadContextParams) -> Result<Self, WhisperError>

### Description
Initializes a new WhisperVadContext using the provided model path and VAD parameters.

### Parameters
- **model_path** (&str) - Required - Path to the whisper model file.
- **params** (WhisperVadContextParams) - Required - Configuration parameters for the VAD context.
```

--------------------------------

### Get Number of Tokens in Segment

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the number of tokens present in this segment.

```rust
pub fn n_tokens(&self) -> c_int
```

--------------------------------

### C++ Equivalent for print_system_info

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.print_system_info.html

This shows the equivalent function signature in C++ for `print_system_info`. It returns a C-style string.

```cpp
const char * whisper_print_system_info()
```

--------------------------------

### print_system_info

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.print_system_info.html

Prints system information. This function is available in Rust and has a C++ equivalent.

```APIDOC
## print_system_info

### Description
Prints system information.

### Signature
```rust
pub fn print_system_info() -> &'static str
```

### C++ Equivalent
```cpp
const char * whisper_print_system_info()
```
```

--------------------------------

### Get segment tokens

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html?search=

Retrieves a specific token by index, with safe and unchecked variants.

```rust
pub fn get_token(&self, token: c_int) -> Option<WhisperToken<'_, '_>>
```

```rust
pub unsafe fn get_token_unchecked(&self, token: c_int) -> WhisperToken<'_, '_>
```

--------------------------------

### Callback Configuration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=

Methods to register callbacks for segment updates, progress tracking, and abort conditions.

```APIDOC
## set_new_segment_callback(new_segment_callback: WhisperNewSegmentCallback)
Set the C callback for new segments. Warning: Cannot be used with DTW.

## set_new_segment_callback_user_data(user_data: *mut c_void)
Set the user data to be passed to the new segment callback.

## set_segment_callback_safe<O, F>(closure: O)
Set the safe callback for segment updates using a closure.

## set_segment_callback_safe_lossy<O, F>(closure: O)
Set the safe callback for segment updates with lossy UTF-8 handling.

## set_progress_callback(progress_callback: WhisperProgressCallback)
Set the C callback for progress updates.

## set_progress_callback_safe<O, F>(closure: O)
Set the safe callback for progress updates (percent).

## set_abort_callback_safe<O, F>(closure: O)
Set the safe callback for abort conditions.
```

--------------------------------

### set_initial_prompt

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Sets the initial prompt text for the model's decoding process.

```APIDOC
## set_initial_prompt(initial_prompt: &str)

### Description
Sets the initial prompt for the model. This text is used as the starting point for the model's decoding. Calling this multiple times will overwrite the previous prompt.

### Parameters
- **initial_prompt** (&str) - Required - A string slice representing the initial prompt text.

### Panics
This method will panic if `initial_prompt` contains a null byte.
```

--------------------------------

### Define WhisperVadSegment struct

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadSegment.html?search=

The structure used to represent a VAD segment with start and end timestamps.

```rust
pub struct WhisperVadSegment {
    pub start: f32,
    pub end: f32,
}
```

--------------------------------

### Get Model Mel Filter Count

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Retrieves the model's mel filter count.

```rust
pub fn model_n_mels(&self) -> c_int
```

```c
int whisper_model_n_mels       (struct whisper_context * ctx);
```

--------------------------------

### Get Segment End Timestamp

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the end time of the segment in centiseconds (10s of milliseconds).

```rust
pub fn end_timestamp(&self) -> i64
```

--------------------------------

### Get Segment Index

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the index of the current segment. This is useful for referencing the segment in other API calls.

```rust
pub fn segment_index(&self) -> c_int
```

--------------------------------

### WhisperContextParameters

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContextParameters.html?search=

Configuration struct for Whisper context initialization.

```APIDOC
## Struct WhisperContextParameters

### Description
Configuration parameters for initializing a Whisper context, including GPU and attention settings.

### Fields
- **use_gpu** (bool) - Use GPU if available.
- **flash_attn** (bool) - Enable flash attention (default false). Note: Cannot be used with DTW.
- **gpu_device** (c_int) - GPU device ID (default 0).
- **dtw_parameters** (DtwParameters) - DTW token level timestamp parameters.

### Methods
- **new() -> Self**: Creates a new instance of WhisperContextParameters.
- **use_gpu(&mut self, use_gpu: bool) -> &mut Self**: Sets the GPU usage flag.
- **flash_attn(&mut self, flash_attn: bool) -> &mut Self**: Sets the flash attention flag.
- **gpu_device(&mut self, gpu_device: c_int) -> &mut Self**: Sets the GPU device ID.
- **dtw_parameters(&mut self, dtw_parameters: DtwParameters) -> &mut Self**: Sets the DTW parameters.
```

--------------------------------

### Get Model Type

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves an integer code representing the overall type or architecture of the Whisper model.

```rust
pub fn model_type(&self) -> c_int
```

--------------------------------

### WhisperVadParams::new

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadParams.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Creates a new instance of WhisperVadParams with default configuration values.

```APIDOC
## pub fn new() -> Self

### Description
Creates a new instance of WhisperVadParams with default configuration values.
```

--------------------------------

### FullParams Configuration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=u32+-%3E+bool

Methods available on the FullParams struct to configure decoding behavior.

```APIDOC
## FullParams Configuration

### new(sampling_strategy: SamplingStrategy) -> FullParams
Creates a new set of parameters for the decoder.

### set_n_threads(n_threads: c_int)
Sets the number of threads to use for decoding. Defaults to min(4, hardware_concurrency).

### set_translate(translate: bool)
Sets whether to translate the output to English. Defaults to false.

### set_language(language: Option<&str>)
Sets the target language. Use "auto" or None for auto-detection. Defaults to "en".

### set_duration_ms(duration_ms: c_int)
Sets the audio duration to process in milliseconds. Defaults to 0.

### set_offset_ms(offset_ms: c_int)
Sets the start offset in milliseconds for decoding. Defaults to 0.

### set_no_timestamps(no_timestamps: bool)
Disables timestamp generation. Defaults to false.
```

--------------------------------

### Get segment text content

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html?search=

Methods to retrieve the segment text as bytes, a UTF-8 string, or a lossy string.

```rust
pub fn to_bytes(&self) -> Result<&'a [u8], WhisperError>
```

```rust
pub fn to_str(&self) -> Result<&'a str, WhisperError>
```

```rust
pub fn to_str_lossy(&self) -> Result<Cow<'a, str>, WhisperError>
```

--------------------------------

### Get Token Data

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html

Retrieves the associated data for a Whisper token. This can provide additional context or information about the token.

```rust
pub fn token_data(&self) -> WhisperTokenData
```

--------------------------------

### WhisperContext::new_with_params

Source: https://docs.rs/whisper-rs/latest/whisper_rs/index.html

Creates a new WhisperContext with specified parameters. This is the initial step for setting up the transcription pipeline.

```APIDOC
## WhisperContext::new_with_params

### Description
Creates a new `WhisperContext` with specified parameters. This is the initial step for setting up the transcription pipeline.

### Method
(Not specified, likely a constructor or associated function)

### Endpoint
(Not applicable, this is an SDK function)

### Parameters
(Specific parameters not detailed in the source, but `WhisperContextParameters` is mentioned as relevant)

### Request Example
(Not applicable)

### Response
#### Success Response
- **WhisperContext**: A new instance of `WhisperContext`.

#### Response Example
(Not applicable)
```

--------------------------------

### C++ Equivalent for Log Callback

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.set_log_callback.html?search=std%3A%3Avec

The corresponding C++ function signature for setting the log callback.

```cpp
void whisper_set_log_callback(whisper_log_callback callback);
```

--------------------------------

### tokenize

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=

Converts text into tokens using the current context.

```APIDOC
## pub fn tokenize(&self, text: &str, max_tokens: usize) -> Result<Vec<WhisperTokenId>, WhisperError>

### Description
Convert the provided text into tokens.

### Arguments
- **text** (&str) - The text to convert.
- **max_tokens** (usize) - The maximum number of tokens.

### Returns
- **Result<Vec<WhisperTokenId>, WhisperError>** - Ok(Vec<WhisperTokenId>) on success, Err(WhisperError) on failure.
```

--------------------------------

### Get Token Probability

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html

Retrieves the probability associated with a Whisper token. This indicates how likely the token is to appear in the given context.

```rust
pub fn token_probability(&self) -> f32
```

--------------------------------

### Decoding Configuration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=

Methods to configure the initial decoding parameters for the whisper model.

```APIDOC
## set_temperature(temperature: f32)
Set initial decoding temperature. Defaults to 0.0.

## set_max_initial_ts(max_initial_ts: f32)
Set max_initial_ts. Defaults to 1.0.

## set_length_penalty(length_penalty: f32)
Set length_penalty. Defaults to -1.0.

## set_temperature_inc(temperature_inc: f32)
Set temperature_inc. Defaults to 0.2.

## set_entropy_thold(entropy_thold: f32)
Set entropy_thold. Defaults to 2.4.

## set_logprob_thold(logprob_thold: f32)
Set logprob_thold. Defaults to -1.0.

## set_no_speech_thold(no_speech_thold: f32)
Set no_speech_thold. Defaults to 0.6.
```

--------------------------------

### Get Token ID

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html

Retrieves the unique identifier for a Whisper token within its segment. This is useful for tracking specific tokens.

```rust
pub fn token_id(&self) -> WhisperTokenId
```

--------------------------------

### Get Specific Token from Segment

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves a `WhisperToken` at a specified index within the segment. Returns `None` if the index is out of bounds.

```rust
pub fn get_token(&self, token: c_int) -> Option<WhisperToken<'_, '_>>
```

--------------------------------

### Get Maximum Language ID

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.get_lang_max_id.html

Returns the ID of the maximum language. This is equivalent to the total number of supported languages minus one.

```rust
pub fn get_lang_max_id() -> i32
```

--------------------------------

### Configure GPU Usage

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContextParameters.html

Sets whether to use the GPU if available for processing.

```rust
pub fn use_gpu(&mut self, use_gpu: bool) -> &mut Self
```

--------------------------------

### WhisperState::full

Source: https://docs.rs/whisper-rs/latest/whisper_rs/index.html

Runs the full transcription pipeline using a `WhisperState`. This is the primary method for obtaining transcription results.

```APIDOC
## WhisperState::full

### Description
Runs the full transcription pipeline using a `WhisperState`. This is the primary method for obtaining transcription results.

### Method
(Not specified, likely a method on `WhisperState`)

### Endpoint
(Not applicable, this is an SDK function)

### Parameters
(Specific parameters for the transcription process, such as audio data and potentially `FullParams`, are implied but not explicitly detailed in the source)

### Request Example
(Not applicable)

### Response
#### Success Response
- **WhisperSegment**: A collection of transcribed segments.

#### Response Example
(Not applicable)
```

--------------------------------

### Get Raw Segment Bytes

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the raw byte representation of the segment's text content. Returns `WhisperError::NullPointer` on failure.

```rust
pub fn to_bytes(&self) -> Result<&'a [u8], WhisperError>
```

--------------------------------

### Rust Function Signature: print_system_info

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.print_system_info.html

This is the Rust signature for the `print_system_info` function. It returns a static string slice containing system information.

```rust
pub fn print_system_info() -> &'static str
```

--------------------------------

### Get Model Type Readable

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Attempts to retrieve a human-readable string describing the model type. This function is exposed but undocumented in the C++ API.

```rust
pub fn model_type_readable_bytes(&self) -> Result<&[u8], WhisperError>
```

--------------------------------

### WhisperContext::new_with_params

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Creates a new WhisperContext from a file path with specified parameters. This is the primary way to load a model from disk.

```APIDOC
## WhisperContext::new_with_params

### Description
Creates a new WhisperContext from a file, with parameters.

### Method
`new_with_params`

### Parameters
#### Path Parameters
- **path** (P: AsRef<Path>) - Required - The path to the model file.
- **parameters** (WhisperContextParameters<'_>) - Required - A parameter struct containing the parameters to use.

#### Returns
- `Result<Self, WhisperError>` - Ok(Self) on success, Err(WhisperError) on failure.

### C++ Equivalent
`struct whisper_context * whisper_init_from_file_with_params_no_state(const char * path_model, struct whisper_context_params params);`
```

--------------------------------

### WhisperStartEncoderCallback

Source: https://docs.rs/whisper-rs/latest/whisper_rs/type.WhisperStartEncoderCallback.html

Represents a callback function that is invoked when the encoder begins. It can either be None or a Some variant containing a C-style function pointer.

```APIDOC
## Type Alias WhisperStartEncoderCallback

```
pub type WhisperStartEncoderCallback = whisper_encoder_begin_callback;
```

## Aliased Type

```
pub enum WhisperStartEncoderCallback {
    None,
    Some(unsafe extern "C" fn(*mut whisper_context, *mut whisper_state, *mut c_void) -> bool),
}
```

### Variants

#### None

No value.

#### Some(unsafe extern "C" fn(*mut whisper_context, *mut whisper_state, *mut c_void) -> bool)

Some value of type `T`.
```

--------------------------------

### Get Specific Token Unchecked

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves a `WhisperToken` at a specified index without bounds checking. This is unsafe and requires the caller to ensure the index is valid.

```rust
pub unsafe fn get_token_unchecked(&self, token: c_int) -> WhisperToken<'_, '_>
```

--------------------------------

### WhisperGrammarElement::new

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperGrammarElement.html?search=u32+-%3E+bool

Creates a new instance of WhisperGrammarElement.

```APIDOC
## pub fn new(element_type: WhisperGrammarElementType, value: u32) -> Self

### Description
Creates a new WhisperGrammarElement with the specified element type and value.

### Parameters
- **element_type** (WhisperGrammarElementType) - Required - The type of the grammar element.
- **value** (u32) - Required - The numeric value associated with the element.
```

--------------------------------

### Get Model Mel Count

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Returns the number of mel spectrogram bins the model uses. This is a parameter related to the audio feature extraction process.

```rust
pub fn model_n_mels(&self) -> c_int
```

--------------------------------

### WhisperState Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html

Methods available on the WhisperState struct for audio processing and model inference.

```APIDOC
## pcm_to_mel

### Description
Convert raw PCM audio (floating point 32 bit) to log mel spectrogram. The resulting spectrogram is stored in the context transparently.

### Parameters
- **pcm** (&[f32]) - Required - The raw PCM audio.
- **threads** (usize) - Required - How many threads to use. Must be at least 1.

### Returns
- **Result<(), WhisperError>** - Ok(()) on success, Err(WhisperError) on failure.

---

## encode

### Description
Run the Whisper encoder on the log mel spectrogram stored inside the provided whisper state.

### Parameters
- **offset** (usize) - Required - The offset of the first frame in the spectrogram.
- **threads** (usize) - Required - How many threads to use. Must be at least 1.

### Returns
- **Result<(), WhisperError>** - Ok(()) on success, Err(WhisperError) on failure.

---

## decode

### Description
Run the Whisper decoder to obtain the logits and probabilities for the next token.

### Parameters
- **tokens** (&[WhisperTokenId]) - Required - The tokens to decode.
- **n_past** (usize) - Required - The number of past tokens to use for the decoding.
- **threads** (usize) - Required - How many threads to use. Must be at least 1.

### Returns
- **Result<(), WhisperError>** - Ok(()) on success, Err(WhisperError) on failure.

---

## lang_detect

### Description
Use mel data at offset_ms to try and auto-detect the spoken language.

### Parameters
- **offset_ms** (usize) - Required - The offset in milliseconds to use for the language detection.
- **threads** (usize) - Required - How many threads to use. Must be at least 1.

### Returns
- **Result<(i32, Vec<f32>), WhisperError>** - Ok((language_id, probabilities)) on success, Err(WhisperError) on failure.
```

--------------------------------

### Callback Registration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.FullParams.html?search=u32+-%3E+bool

Methods to register callbacks for monitoring transcription progress and handling segments.

```APIDOC
## set_segment_callback_safe<O, F>(closure: O)
Registers a safe closure to be called on segment updates.

## set_segment_callback_safe_lossy<O, F>(closure: O)
Registers a safe closure for segment updates with lossy handling of invalid UTF-8 characters.

## set_progress_callback_safe<O, F>(closure: O)
Registers a safe closure to receive progress updates as a percentage (i32).

## set_abort_callback_safe<O, F>(closure: O)
Registers a safe closure that returns a boolean to determine if the transcription should abort.
```

--------------------------------

### Get Full Language String by ID

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.get_lang_str_full.html

Retrieves the full string name of a language using its integer ID. Returns None if the ID is not found.

```rust
pub fn get_lang_str_full(id: i32) -> Option<&'static str>
```

--------------------------------

### Get Raw Token Bytes

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html

Retrieves the raw byte representation of a Whisper token. This is useful for languages where tokens may not align with UTF-8 character boundaries.

```rust
pub fn to_bytes(&self) -> Result<&'b [u8], WhisperError>
```

--------------------------------

### Retrieve Token as String

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html?search=std%3A%3Avec

Retrieves the UTF-8 validated text of the token.

```rust
pub fn to_str(&self) -> Result<&'b str, WhisperError>
```

--------------------------------

### Get Model Text Layer Count

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the number of layers in the text processing part of the model. This indicates the depth of the text generation or understanding capabilities.

```rust
pub fn model_n_text_layer(&self) -> c_int
```

--------------------------------

### Query Model Metadata

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=std%3A%3Avec

Methods to retrieve various configuration and structural properties of the loaded model.

```rust
pub fn n_vocab(&self) -> c_int
```

```cpp
int whisper_n_vocab        (struct whisper_context * ctx)
```

```rust
pub fn n_text_ctx(&self) -> c_int
```

```cpp
int whisper_n_text_ctx     (struct whisper_context * ctx);
```

```rust
pub fn n_audio_ctx(&self) -> c_int
```

```cpp
int whisper_n_audio_ctx     (struct whisper_context * ctx);
```

```rust
pub fn is_multilingual(&self) -> bool
```

```cpp
int whisper_is_multilingual(struct whisper_context * ctx)
```

```rust
pub fn model_n_vocab(&self) -> c_int
```

```cpp
int whisper_model_n_vocab      (struct whisper_context * ctx);
```

```rust
pub fn model_n_audio_ctx(&self) -> c_int
```

```cpp
int whisper_model_n_audio_ctx    (struct whisper_context * ctx)
```

```rust
pub fn model_n_audio_state(&self) -> c_int
```

```cpp
int whisper_model_n_audio_state(struct whisper_context * ctx);
```

```rust
pub fn model_n_audio_head(&self) -> c_int
```

```cpp
int whisper_model_n_audio_head (struct whisper_context * ctx);
```

```rust
pub fn model_n_audio_layer(&self) -> c_int
```

```cpp
int whisper_model_n_audio_layer(struct whisper_context * ctx);
```

```rust
pub fn model_n_text_ctx(&self) -> c_int
```

```cpp
int whisper_model_n_text_ctx     (struct whisper_context * ctx)
```

```rust
pub fn model_n_text_state(&self) -> c_int
```

```cpp
int whisper_model_n_text_state (struct whisper_context * ctx);
```

```rust
pub fn model_n_text_head(&self) -> c_int
```

```cpp
int whisper_model_n_text_head  (struct whisper_context * ctx);
```

```rust
pub fn model_n_text_layer(&self) -> c_int
```

```cpp
int whisper_model_n_text_layer (struct whisper_context * ctx);
```

```rust
pub fn model_n_mels(&self) -> c_int
```

```cpp
int whisper_model_n_mels       (struct whisper_context * ctx);
```

```rust
pub fn model_ftype(&self) -> c_int
```

```cpp
int whisper_model_ftype          (struct whisper_context * ctx);
```

```rust
pub fn model_type(&self) -> c_int
```

```cpp
int whisper_model_type         (struct whisper_context * ctx);
```

--------------------------------

### Function Signature for convert_stereo_to_mono_audio

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.convert_stereo_to_mono_audio.html?search=

The function signature defines the input stereo samples and the mutable output buffer for mono samples.

```rust
pub fn convert_stereo_to_mono_audio(
    input: &[f32],
    output: &mut [f32],
) -> Result<(), WhisperError>
```

--------------------------------

### Run Whisper Encoder

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Executes the encoder on the stored log mel spectrogram.

```rust
pub fn encode( &mut self, offset: usize, threads: usize, ) -> Result<(), WhisperError>
```

```c
int whisper_encode(struct whisper_context * ctx, int offset, int n_threads)
```

--------------------------------

### print_timings

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Prints performance statistics to stderr.

```APIDOC
## pub fn print_timings(&self)

### Description
Prints detailed performance statistics and timing information related to the Whisper context operations to the standard error stream (stderr).

### C++ Equivalent
`void whisper_print_timings(struct whisper_context * ctx);`
```

--------------------------------

### Get Token Text (UTF-8 Validated)

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html

Retrieves the text of a Whisper token, ensuring it is valid UTF-8. Returns an error if the token data is invalid or a null pointer.

```rust
pub fn to_str(&self) -> Result<&'b str, WhisperError>
```

--------------------------------

### Check Multilingual Support

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Determines if the model supports multiple languages.

```rust
pub fn is_multilingual(&self) -> bool
```

```c
int whisper_is_multilingual(struct whisper_context * ctx)
```

--------------------------------

### WhisperVadParams Configuration Methods

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadParams.html

Methods to configure VAD behavior including thresholds, durations, and padding.

```APIDOC
## Configuration Methods

### pub fn set_threshold(&mut self, threshold: f32)
Sets the probability threshold (0.0 to 1.0) to consider a segment as speech. Defaults to 0.5.

### pub fn set_min_speech_duration(&mut self, min_speech_duration: c_int)
Sets the minimum duration for a valid speech segment in milliseconds. Defaults to 250ms.

### pub fn set_min_silence_duration(&mut self, min_silence_duration: c_int)
Sets the minimum silence duration to consider speech as ended in milliseconds. Defaults to 100ms.

### pub fn set_max_speech_duration(&mut self, max_speech_duration: f32)
Sets the maximum duration of a speech segment before forcing a new segment. Defaults to f32::MAX.

### pub fn set_speech_pad(&mut self, speech_pad: c_int)
Sets the padding added before and after speech segments in milliseconds. Defaults to 30ms.

### pub fn set_samples_overlap(&mut self, samples_overlap: f32)
Sets the amount of audio to extend from each speech segment into the next one in seconds. Defaults to 0.1s.
```

--------------------------------

### Get Segment Text as UTF-8 String

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the segment's text content as a UTF-8 validated string. Returns `WhisperError::NullPointer` or `WhisperError::InvalidUtf8` on failure.

```rust
pub fn to_str(&self) -> Result<&'a str, WhisperError>
```

--------------------------------

### WhisperVadContextParams::new

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadContextParams.html?search=

Creates a new instance of WhisperVadContextParams with default settings.

```APIDOC
## pub fn new() -> Self

### Description
Creates a new instance of WhisperVadContextParams.

### Signature
`pub fn new() -> Self`
```

--------------------------------

### Get Model Text State Size

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the size of the internal text state maintained by the model. This relates to the model's memory for text generation or processing.

```rust
pub fn model_n_text_state(&self) -> c_int
```

--------------------------------

### Run Whisper Decoder

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

Decodes tokens to obtain logits and probabilities for the next token.

```rust
pub fn decode( &mut self, tokens: &[WhisperTokenId], n_past: usize, threads: usize, ) -> Result<(), WhisperError>
```

```c
int whisper_decode(struct whisper_context * ctx, const whisper_token * tokens, int n_tokens, int n_past, int n_threads)
```

--------------------------------

### Get Model Audio Layer Count

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Retrieves the number of layers in the audio processing part of the model. More layers generally mean a more complex and potentially more capable model.

```rust
pub fn model_n_audio_layer(&self) -> c_int
```

--------------------------------

### convert_stereo_to_mono_audio

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.convert_stereo_to_mono_audio.html

Converts 32-bit floating point stereo PCM audio to 32-bit floating point mono PCM audio. It takes an input slice of stereo samples and an output mutable slice for mono samples. Errors can occur if the input sample length is odd or if there's a mismatch between input and output lengths.

```APIDOC
## Function convert_stereo_to_mono_audio

### Description
Converts 32-bit floating point stereo PCM audio to 32-bit floating point mono PCM audio.

### Arguments
* `input` - The array of 32-bit floating point stereo PCM audio samples.
* `output` - An output place to write all the mono samples.

### Errors
* if `samples.len()` is odd (`WhisperError::HalfSampleMissing`)
* if `input.len() / 2 < samples.len()` (`WhisperError::InputOutputLengthMismatch`)

### Returns
A vector of 32-bit floating point mono PCM audio samples.

### Examples
```rust
let samples = [0.0f32; 1024];
let mut mono_samples = [0.0f32; 512];
convert_stereo_to_mono_audio(&samples, &mut mono_samples).expect("should be no half samples missing");
```
```

--------------------------------

### Get Language ID

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.get_lang_id.html

Returns the integer ID for a given language string. Returns None if the language is not found. Panics if the language string contains a null byte.

```rust
pub fn get_lang_id(lang: &str) -> Option<c_int>
```

--------------------------------

### WhisperVadContext::segments_from_samples

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperVadContext.html

Runs the full VAD pipeline on audio samples.

```APIDOC
## pub fn segments_from_samples(&mut self, params: WhisperVadParams, samples: &[f32]) -> Result<WhisperVadSegments, WhisperError>

### Description
Executes the entire VAD pipeline, including speech detection and segment extraction, in one call.

### Parameters
- **params** (WhisperVadParams) - Required - VAD parameters.
- **samples** (&[f32]) - Required - Audio samples to process.
```

--------------------------------

### Get Segment Text as Lossy String

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperSegment.html

Retrieves the segment's text content, replacing invalid UTF-8 sequences with the replacement character. Returns `WhisperError::NullPointer` on failure.

```rust
pub fn to_str_lossy(&self) -> Result<Cow<'a, str>, WhisperError>
```

--------------------------------

### decode

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=

Runs the Whisper decoder to obtain logits and probabilities for the next token.

```APIDOC
## pub fn decode(&mut self, tokens: &[WhisperTokenId], n_past: usize, threads: usize) -> Result<(), WhisperError>

### Description
Run the Whisper decoder to obtain the logits and probabilities for the next token. Make sure to call WhisperState::encode first.

### Parameters
- **tokens** (&[WhisperTokenId]) - Required - The tokens to decode.
- **n_past** (usize) - Required - The number of past tokens to use for the decoding.
- **threads** (usize) - Required - How many threads to use. Must be at least 1.
```

--------------------------------

### WhisperContextParameters::new

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContextParameters.html?search=u32+-%3E+bool

Creates a new instance of WhisperContextParameters with default values.

```APIDOC
## pub fn new() -> Self

### Description
Creates a new instance of WhisperContextParameters with default settings.

### Signature
`pub fn new() -> Self`
```

--------------------------------

### Get Model Text Head Count

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContext.html

Returns the number of attention heads in the text processing layers of the model. This affects how the model weighs different tokens during text generation.

```rust
pub fn model_n_text_head(&self) -> c_int
```

--------------------------------

### Function Signature for convert_integer_to_float_audio

Source: https://docs.rs/whisper-rs/latest/whisper_rs/fn.convert_integer_to_float_audio.html?search=Option%3CT%3E%2C+%28T+-%3E+U%29+-%3E+Option%3CU%3E

The function signature defines the input slice of 16-bit integers and the mutable output slice of 32-bit floats, returning a Result.

```rust
pub fn convert_integer_to_float_audio(
    samples: &[i16],
    output: &mut [f32],
) -> Result<(), WhisperError>
```

--------------------------------

### Get Token Text (Lossy UTF-8)

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperToken.html

Retrieves the text of a Whisper token, replacing invalid UTF-8 sequences with the replacement character. This method is useful for handling potentially malformed text.

```rust
pub fn to_str_lossy(&self) -> Result<Cow<'b, str>, WhisperError>
```

--------------------------------

### pcm_to_mel

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperState.html?search=

Converts raw PCM audio to a log mel spectrogram stored within the state.

```APIDOC
## pub fn pcm_to_mel(&mut self, pcm: &[f32], threads: usize) -> Result<(), WhisperError>

### Description
Convert raw PCM audio (floating point 32 bit) to log mel spectrogram. The resulting spectrogram is stored in the context transparently.

### Parameters
- **pcm** (&[f32]) - Required - The raw PCM audio.
- **threads** (usize) - Required - How many threads to use. Must be at least 1.
```

--------------------------------

### Enable Flash Attention

Source: https://docs.rs/whisper-rs/latest/whisper_rs/struct.WhisperContextParameters.html

Enables flash attention for potentially faster processing. Note: Cannot be used with DTW; DTW will be disabled if flash_attn is true.

```rust
pub fn flash_attn(&mut self, flash_attn: bool) -> &mut Self
```
Tokens:
Raw