Package io.github.ggerganov.whispercpp
Interface WhisperCppJnaLibrary
- All Superinterfaces:
com.sun.jna.Library
public interface WhisperCppJnaLibrary
extends com.sun.jna.Library
-
Nested Class Summary
Nested classes/interfaces inherited from interface com.sun.jna.Library
com.sun.jna.Library.Handler -
Field Summary
FieldsFields inherited from interface com.sun.jna.Library
OPTION_ALLOW_OBJECTS, OPTION_CALLING_CONVENTION, OPTION_CLASSLOADER, OPTION_FUNCTION_MAPPER, OPTION_INVOCATION_MAPPER, OPTION_OPEN_FLAGS, OPTION_STRING_ENCODING, OPTION_STRUCTURE_ALIGNMENT, OPTION_SYMBOL_PROVIDER, OPTION_TYPE_MAPPER -
Method Summary
Modifier and TypeMethodDescriptionintwhisper_bench_ggml_mul_mat(int nThreads) Benchmark function for ggml_mul_mat.whisper_bench_ggml_mul_mat_str(int nThreads) Benchmark function for ggml_mul_mat as a string.intwhisper_bench_memcpy(int nThreads) Benchmark function for memcpy.whisper_bench_memcpy_str(int nThreads) Benchmark function for memcpy as a string.intwhisper_decode(com.sun.jna.Pointer ctx, com.sun.jna.Pointer tokens, int n_tokens, int n_past, int n_threads) Run the Whisper decoder to obtain the logits and probabilities for the next token.intwhisper_decode_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, com.sun.jna.Pointer tokens, int n_tokens, int n_past, int n_threads) intwhisper_encode(com.sun.jna.Pointer ctx, int offset, int n_threads) Run the Whisper encoder on the log mel spectrogram stored inside the default state in the provided whisper context.intwhisper_encode_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, int offset, int n_threads) voidwhisper_free(com.sun.jna.Pointer ctx) Free all allocated memory associated with the Whisper context.voidwhisper_free_params(com.sun.jna.Pointer params) voidwhisper_free_state(com.sun.jna.Pointer state) Free all allocated memory associated with the Whisper state.intwhisper_full(com.sun.jna.Pointer ctx, WhisperFullParams params, float[] samples, int n_samples) Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text Not thread safe for same context Uses the specified decoding strategy to obtain the text.com.sun.jna.Pointerwhisper_full_default_params_by_ref(int strategy) Provides default params which can be used with `whisper_full()` etc.longwhisper_full_get_segment_t0(com.sun.jna.Pointer ctx, int i_segment) Get the start time of the specified segment.longwhisper_full_get_segment_t0_from_state(com.sun.jna.Pointer state, int i_segment) Get the start time of the specified segment from the state.longwhisper_full_get_segment_t1(com.sun.jna.Pointer ctx, int i_segment) Get the end time of the specified segment.longwhisper_full_get_segment_t1_from_state(com.sun.jna.Pointer state, int i_segment) Get the end time of the specified segment from the state.whisper_full_get_segment_text(com.sun.jna.Pointer ctx, int i_segment) Get the text of the specified segment.whisper_full_get_segment_text_from_state(com.sun.jna.Pointer state, int i_segment) Get the text of the specified segment from the state.whisper_full_get_token_data(com.sun.jna.Pointer ctx, int i_segment, int i_token) Get token data for the specified token in the specified segment.whisper_full_get_token_data_from_state(com.sun.jna.Pointer state, int i_segment, int i_token) Get token data for the specified token in the specified segment from the state.intwhisper_full_get_token_id(com.sun.jna.Pointer ctx, int i_segment, int i_token) Get the token ID of the specified token in the specified segment.intwhisper_full_get_token_id_from_state(com.sun.jna.Pointer state, int i_segment, int i_token) Get the token ID of the specified token in the specified segment from the state.floatwhisper_full_get_token_p(com.sun.jna.Pointer ctx, int i_segment, int i_token) Get the probability of the specified token in the specified segment.floatwhisper_full_get_token_p_from_state(com.sun.jna.Pointer state, int i_segment, int i_token) Get the probability of the specified token in the specified segment from the state.whisper_full_get_token_text(com.sun.jna.Pointer ctx, int i_segment, int i_token) Get the token text of the specified token in the specified segment.whisper_full_get_token_text_from_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, int i_segment, int i_token) Get the token text of the specified token in the specified segment from the state.intwhisper_full_lang_id(com.sun.jna.Pointer ctx) Language id associated with the context's default state.intwhisper_full_lang_id_from_state(com.sun.jna.Pointer state) Language id associated with the provided stateintwhisper_full_n_segments(com.sun.jna.Pointer ctx) Number of generated text segments.intwhisper_full_n_segments_from_state(com.sun.jna.Pointer state) intwhisper_full_n_tokens(com.sun.jna.Pointer ctx, int i_segment) Get the number of tokens in the specified segment.intwhisper_full_n_tokens_from_state(com.sun.jna.Pointer state, int i_segment) Get the number of tokens in the specified segment from the state.intwhisper_full_parallel(com.sun.jna.Pointer ctx, WhisperFullParams params, float[] samples, int n_samples, int n_processors) intwhisper_full_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, WhisperFullParams params, float[] samples, int n_samples) float[]whisper_get_logits(com.sun.jna.Pointer ctx) Token logits obtained from the last call to whisper_decode().float[]whisper_get_logits_from_state(com.sun.jna.Pointer state) com.sun.jna.Pointerwhisper_init(WhisperModelLoader loader) Allocate (almost) all memory needed for the model using a model loader.com.sun.jna.Pointerwhisper_init_from_buffer(com.sun.jna.Pointer buffer, int buffer_size) Allocate (almost) all memory needed for the model by loading from a buffer.com.sun.jna.Pointerwhisper_init_from_buffer_no_state(com.sun.jna.Pointer buffer, int buffer_size) Allocate (almost) all memory needed for the model by loading from a buffer without allocating the state.com.sun.jna.Pointerwhisper_init_from_file(String path_model) Allocate (almost) all memory needed for the model by loading from a file.com.sun.jna.Pointerwhisper_init_from_file_no_state(String path_model) Allocate (almost) all memory needed for the model by loading from a file without allocating the state.com.sun.jna.PointerAllocate (almost) all memory needed for the model using a model loader without allocating the state.com.sun.jna.Pointerwhisper_init_state(com.sun.jna.Pointer ctx) Allocate memory for the Whisper state.intwhisper_is_multilingual(com.sun.jna.Pointer ctx) intwhisper_lang_auto_detect(com.sun.jna.Pointer ctx, int offset_ms, int n_threads, float[] lang_probs) Use mel data at offset_ms to try and auto-detect the spoken language.intwhisper_lang_auto_detect_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, int offset_ms, int n_threads, float[] lang_probs) intwhisper_lang_id(String lang) intLargest language id (i.e.whisper_lang_str(int id) intwhisper_model_ftype(com.sun.jna.Pointer ctx) intwhisper_model_n_audio_ctx(com.sun.jna.Pointer ctx) intwhisper_model_n_audio_head(com.sun.jna.Pointer ctx) intwhisper_model_n_audio_layer(com.sun.jna.Pointer ctx) intwhisper_model_n_audio_state(com.sun.jna.Pointer ctx) intwhisper_model_n_mels(com.sun.jna.Pointer ctx) intwhisper_model_n_text_ctx(com.sun.jna.Pointer ctx) intwhisper_model_n_text_head(com.sun.jna.Pointer ctx) intwhisper_model_n_text_layer(com.sun.jna.Pointer ctx) intwhisper_model_n_text_state(com.sun.jna.Pointer ctx) intwhisper_model_n_vocab(com.sun.jna.Pointer ctx) intwhisper_model_type(com.sun.jna.Pointer ctx) whisper_model_type_readable(com.sun.jna.Pointer ctx) intwhisper_n_audio_ctx(com.sun.jna.Pointer ctx) intwhisper_n_len(com.sun.jna.Pointer ctx) intwhisper_n_len_from_state(com.sun.jna.Pointer state) intwhisper_n_text_ctx(com.sun.jna.Pointer ctx) intwhisper_n_vocab(com.sun.jna.Pointer ctx) intwhisper_pcm_to_mel(com.sun.jna.Pointer ctx, float[] samples, int n_samples, int n_threads) Convert RAW PCM audio to log mel spectrogram.intwhisper_pcm_to_mel_phase_vocoder(com.sun.jna.Pointer ctx, float[] samples, int n_samples, int n_threads) Convert RAW PCM audio to log mel spectrogram but applies a Phase Vocoder to speed up the audio x2.intwhisper_pcm_to_mel_phase_vocoder_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, float[] samples, int n_samples, int n_threads) intwhisper_pcm_to_mel_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, float[] samples, int n_samples, int n_threads) voidwhisper_print_timings(com.sun.jna.Pointer ctx) voidwhisper_reset_timings(com.sun.jna.Pointer ctx) intwhisper_set_mel(com.sun.jna.Pointer ctx, float[] data, int n_len, int n_mel) This can be used to set a custom log mel spectrogram inside the default state of the provided whisper context.intwhisper_set_mel_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, float[] data, int n_len, int n_mel) intwhisper_token_beg(com.sun.jna.Pointer ctx) intwhisper_token_eot(com.sun.jna.Pointer ctx) intwhisper_token_lang(com.sun.jna.Pointer ctx, int lang_id) intwhisper_token_not(com.sun.jna.Pointer ctx) intwhisper_token_prev(com.sun.jna.Pointer ctx) intwhisper_token_solm(com.sun.jna.Pointer ctx) intwhisper_token_sot(com.sun.jna.Pointer ctx) whisper_token_to_str(com.sun.jna.Pointer ctx, int token) intwhisper_token_transcribe(com.sun.jna.Pointer ctx) intwhisper_token_translate(com.sun.jna.Pointer ctx) intwhisper_tokenize(com.sun.jna.Pointer ctx, String text, com.sun.jna.Pointer tokens, int n_max_tokens) Convert the provided text into tokens.
-
Field Details
-
instance
-
-
Method Details
-
whisper_print_system_info
String whisper_print_system_info() -
whisper_init_from_file
Allocate (almost) all memory needed for the model by loading from a file.- Parameters:
path_model- Path to the model file- Returns:
- Whisper context on success, null on failure
-
whisper_init_from_buffer
com.sun.jna.Pointer whisper_init_from_buffer(com.sun.jna.Pointer buffer, int buffer_size) Allocate (almost) all memory needed for the model by loading from a buffer.- Parameters:
buffer- Model bufferbuffer_size- Size of the model buffer- Returns:
- Whisper context on success, null on failure
-
whisper_init
Allocate (almost) all memory needed for the model using a model loader.- Parameters:
loader- Model loader- Returns:
- Whisper context on success, null on failure
-
whisper_init_from_file_no_state
Allocate (almost) all memory needed for the model by loading from a file without allocating the state.- Parameters:
path_model- Path to the model file- Returns:
- Whisper context on success, null on failure
-
whisper_init_from_buffer_no_state
com.sun.jna.Pointer whisper_init_from_buffer_no_state(com.sun.jna.Pointer buffer, int buffer_size) Allocate (almost) all memory needed for the model by loading from a buffer without allocating the state.- Parameters:
buffer- Model bufferbuffer_size- Size of the model buffer- Returns:
- Whisper context on success, null on failure
-
whisper_init_no_state
Allocate (almost) all memory needed for the model using a model loader without allocating the state.- Parameters:
loader- Model loader- Returns:
- Whisper context on success, null on failure
-
whisper_init_state
com.sun.jna.Pointer whisper_init_state(com.sun.jna.Pointer ctx) Allocate memory for the Whisper state.- Parameters:
ctx- Whisper context- Returns:
- Whisper state on success, null on failure
-
whisper_free
void whisper_free(com.sun.jna.Pointer ctx) Free all allocated memory associated with the Whisper context.- Parameters:
ctx- Whisper context
-
whisper_free_state
void whisper_free_state(com.sun.jna.Pointer state) Free all allocated memory associated with the Whisper state.- Parameters:
state- Whisper state
-
whisper_pcm_to_mel
int whisper_pcm_to_mel(com.sun.jna.Pointer ctx, float[] samples, int n_samples, int n_threads) Convert RAW PCM audio to log mel spectrogram. The resulting spectrogram is stored inside the default state of the provided whisper context.- Parameters:
ctx- - Pointer to a WhisperContext- Returns:
- 0 on success
-
whisper_pcm_to_mel_with_state
int whisper_pcm_to_mel_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, float[] samples, int n_samples, int n_threads) - Parameters:
ctx- Pointer to a WhisperContextstate- Pointer to WhisperStaten_samples-n_threads-- Returns:
- 0 on success
-
whisper_set_mel
int whisper_set_mel(com.sun.jna.Pointer ctx, float[] data, int n_len, int n_mel) This can be used to set a custom log mel spectrogram inside the default state of the provided whisper context. Use this instead of whisper_pcm_to_mel() if you want to provide your own log mel spectrogram. n_mel must be 80- Returns:
- 0 on success
-
whisper_set_mel_with_state
int whisper_set_mel_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, float[] data, int n_len, int n_mel) -
whisper_encode
int whisper_encode(com.sun.jna.Pointer ctx, int offset, int n_threads) Run the Whisper encoder on the log mel spectrogram stored inside the default state in the provided whisper context. Make sure to call whisper_pcm_to_mel() or whisper_set_mel() first. Offset can be used to specify the offset of the first frame in the spectrogram.- Returns:
- 0 on success
-
whisper_encode_with_state
int whisper_encode_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, int offset, int n_threads) -
whisper_decode
int whisper_decode(com.sun.jna.Pointer ctx, com.sun.jna.Pointer tokens, int n_tokens, int n_past, int n_threads) Run the Whisper decoder to obtain the logits and probabilities for the next token. Make sure to call whisper_encode() first. tokens + n_tokens is the provided context for the decoder. n_past is the number of tokens to use from previous decoder calls. Returns 0 on success TODO: add support for multiple decoders -
whisper_decode_with_state
int whisper_decode_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, com.sun.jna.Pointer tokens, int n_tokens, int n_past, int n_threads) - Parameters:
ctx-state-tokens- Pointer to int tokensn_tokens-n_past-n_threads-- Returns:
-
whisper_tokenize
int whisper_tokenize(com.sun.jna.Pointer ctx, String text, com.sun.jna.Pointer tokens, int n_max_tokens) Convert the provided text into tokens. The tokens pointer must be large enough to hold the resulting tokens. Returns the number of tokens on success, no more than n_max_tokens Returns -1 on failure TODO: not sure if correct -
whisper_lang_max_id
int whisper_lang_max_id()Largest language id (i.e. number of available languages - 1) -
whisper_lang_id
- Returns:
- the id of the specified language, returns -1 if not found. Examples: "de" -> 2 "german" -> 2
-
whisper_lang_str
- Returns:
- the short string of the specified language id (e.g. 2 -> "de"), returns nullptr if not found
-
whisper_lang_auto_detect
int whisper_lang_auto_detect(com.sun.jna.Pointer ctx, int offset_ms, int n_threads, float[] lang_probs) Use mel data at offset_ms to try and auto-detect the spoken language. Make sure to call whisper_pcm_to_mel() or whisper_set_mel() first Returns the top language id or negative on failure If not null, fills the lang_probs array with the probabilities of all languages The array must be whisper_lang_max_id() + 1 in size ref: https://github.com/openai/whisper/blob/main/whisper/decoding.py#L18-L69 -
whisper_lang_auto_detect_with_state
int whisper_lang_auto_detect_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, int offset_ms, int n_threads, float[] lang_probs) -
whisper_n_len
int whisper_n_len(com.sun.jna.Pointer ctx) -
whisper_n_len_from_state
int whisper_n_len_from_state(com.sun.jna.Pointer state) -
whisper_n_vocab
int whisper_n_vocab(com.sun.jna.Pointer ctx) -
whisper_n_text_ctx
int whisper_n_text_ctx(com.sun.jna.Pointer ctx) -
whisper_n_audio_ctx
int whisper_n_audio_ctx(com.sun.jna.Pointer ctx) -
whisper_is_multilingual
int whisper_is_multilingual(com.sun.jna.Pointer ctx) -
whisper_model_n_vocab
int whisper_model_n_vocab(com.sun.jna.Pointer ctx) -
whisper_model_n_audio_ctx
int whisper_model_n_audio_ctx(com.sun.jna.Pointer ctx) -
whisper_model_n_audio_state
int whisper_model_n_audio_state(com.sun.jna.Pointer ctx) -
whisper_model_n_audio_head
int whisper_model_n_audio_head(com.sun.jna.Pointer ctx) -
whisper_model_n_audio_layer
int whisper_model_n_audio_layer(com.sun.jna.Pointer ctx) -
whisper_model_n_text_ctx
int whisper_model_n_text_ctx(com.sun.jna.Pointer ctx) -
whisper_model_n_text_state
int whisper_model_n_text_state(com.sun.jna.Pointer ctx) -
whisper_model_n_text_head
int whisper_model_n_text_head(com.sun.jna.Pointer ctx) -
whisper_model_n_text_layer
int whisper_model_n_text_layer(com.sun.jna.Pointer ctx) -
whisper_model_n_mels
int whisper_model_n_mels(com.sun.jna.Pointer ctx) -
whisper_model_ftype
int whisper_model_ftype(com.sun.jna.Pointer ctx) -
whisper_model_type
int whisper_model_type(com.sun.jna.Pointer ctx) -
whisper_get_logits
float[] whisper_get_logits(com.sun.jna.Pointer ctx) Token logits obtained from the last call to whisper_decode(). The logits for the last token are stored in the last row Rows: n_tokens Cols: n_vocab -
whisper_get_logits_from_state
float[] whisper_get_logits_from_state(com.sun.jna.Pointer state) -
whisper_token_to_str
-
whisper_model_type_readable
-
whisper_token_eot
int whisper_token_eot(com.sun.jna.Pointer ctx) -
whisper_token_sot
int whisper_token_sot(com.sun.jna.Pointer ctx) -
whisper_token_prev
int whisper_token_prev(com.sun.jna.Pointer ctx) -
whisper_token_solm
int whisper_token_solm(com.sun.jna.Pointer ctx) -
whisper_token_not
int whisper_token_not(com.sun.jna.Pointer ctx) -
whisper_token_beg
int whisper_token_beg(com.sun.jna.Pointer ctx) -
whisper_token_lang
int whisper_token_lang(com.sun.jna.Pointer ctx, int lang_id) -
whisper_token_translate
int whisper_token_translate(com.sun.jna.Pointer ctx) -
whisper_token_transcribe
int whisper_token_transcribe(com.sun.jna.Pointer ctx) -
whisper_print_timings
void whisper_print_timings(com.sun.jna.Pointer ctx) -
whisper_reset_timings
void whisper_reset_timings(com.sun.jna.Pointer ctx) -
whisper_full_default_params_by_ref
com.sun.jna.Pointer whisper_full_default_params_by_ref(int strategy) Provides default params which can be used with `whisper_full()` etc. Because this function allocates memory for the params, the caller must call either: - call `whisper_free_params()` - `Native.free(Pointer.nativeValue(pointer));`- Parameters:
strategy- - WhisperSamplingStrategy.value
-
whisper_free_params
void whisper_free_params(com.sun.jna.Pointer params) -
whisper_full
Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text Not thread safe for same context Uses the specified decoding strategy to obtain the text. -
whisper_full_with_state
int whisper_full_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, WhisperFullParams params, float[] samples, int n_samples) -
whisper_full_parallel
int whisper_full_parallel(com.sun.jna.Pointer ctx, WhisperFullParams params, float[] samples, int n_samples, int n_processors) -
whisper_full_n_segments
int whisper_full_n_segments(com.sun.jna.Pointer ctx) Number of generated text segments. A segment can be a few words, a sentence, or even a paragraph.- Parameters:
ctx- Pointer to WhisperContext
-
whisper_full_n_segments_from_state
int whisper_full_n_segments_from_state(com.sun.jna.Pointer state) - Parameters:
state- Pointer to WhisperState
-
whisper_full_lang_id
int whisper_full_lang_id(com.sun.jna.Pointer ctx) Language id associated with the context's default state.- Parameters:
ctx- Pointer to WhisperContext
-
whisper_full_lang_id_from_state
int whisper_full_lang_id_from_state(com.sun.jna.Pointer state) Language id associated with the provided state -
whisper_pcm_to_mel_phase_vocoder
int whisper_pcm_to_mel_phase_vocoder(com.sun.jna.Pointer ctx, float[] samples, int n_samples, int n_threads) Convert RAW PCM audio to log mel spectrogram but applies a Phase Vocoder to speed up the audio x2. The resulting spectrogram is stored inside the default state of the provided whisper context.- Returns:
- 0 on success
-
whisper_pcm_to_mel_phase_vocoder_with_state
int whisper_pcm_to_mel_phase_vocoder_with_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, float[] samples, int n_samples, int n_threads) -
whisper_full_get_segment_t0
long whisper_full_get_segment_t0(com.sun.jna.Pointer ctx, int i_segment) Get the start time of the specified segment. -
whisper_full_get_segment_t0_from_state
long whisper_full_get_segment_t0_from_state(com.sun.jna.Pointer state, int i_segment) Get the start time of the specified segment from the state. -
whisper_full_get_segment_t1
long whisper_full_get_segment_t1(com.sun.jna.Pointer ctx, int i_segment) Get the end time of the specified segment. -
whisper_full_get_segment_t1_from_state
long whisper_full_get_segment_t1_from_state(com.sun.jna.Pointer state, int i_segment) Get the end time of the specified segment from the state. -
whisper_full_get_segment_text
Get the text of the specified segment. -
whisper_full_get_segment_text_from_state
Get the text of the specified segment from the state. -
whisper_full_n_tokens
int whisper_full_n_tokens(com.sun.jna.Pointer ctx, int i_segment) Get the number of tokens in the specified segment. -
whisper_full_n_tokens_from_state
int whisper_full_n_tokens_from_state(com.sun.jna.Pointer state, int i_segment) Get the number of tokens in the specified segment from the state. -
whisper_full_get_token_text
Get the token text of the specified token in the specified segment. -
whisper_full_get_token_text_from_state
String whisper_full_get_token_text_from_state(com.sun.jna.Pointer ctx, com.sun.jna.Pointer state, int i_segment, int i_token) Get the token text of the specified token in the specified segment from the state. -
whisper_full_get_token_id
int whisper_full_get_token_id(com.sun.jna.Pointer ctx, int i_segment, int i_token) Get the token ID of the specified token in the specified segment. -
whisper_full_get_token_id_from_state
int whisper_full_get_token_id_from_state(com.sun.jna.Pointer state, int i_segment, int i_token) Get the token ID of the specified token in the specified segment from the state. -
whisper_full_get_token_data
Get token data for the specified token in the specified segment. -
whisper_full_get_token_data_from_state
WhisperTokenData whisper_full_get_token_data_from_state(com.sun.jna.Pointer state, int i_segment, int i_token) Get token data for the specified token in the specified segment from the state. -
whisper_full_get_token_p
float whisper_full_get_token_p(com.sun.jna.Pointer ctx, int i_segment, int i_token) Get the probability of the specified token in the specified segment. -
whisper_full_get_token_p_from_state
float whisper_full_get_token_p_from_state(com.sun.jna.Pointer state, int i_segment, int i_token) Get the probability of the specified token in the specified segment from the state. -
whisper_bench_memcpy
int whisper_bench_memcpy(int nThreads) Benchmark function for memcpy.- Parameters:
nThreads- Number of threads to use for the benchmark.- Returns:
- The result of the benchmark.
-
whisper_bench_memcpy_str
Benchmark function for memcpy as a string.- Parameters:
nThreads- Number of threads to use for the benchmark.- Returns:
- The result of the benchmark as a string.
-
whisper_bench_ggml_mul_mat
int whisper_bench_ggml_mul_mat(int nThreads) Benchmark function for ggml_mul_mat.- Parameters:
nThreads- Number of threads to use for the benchmark.- Returns:
- The result of the benchmark.
-
whisper_bench_ggml_mul_mat_str
Benchmark function for ggml_mul_mat as a string.- Parameters:
nThreads- Number of threads to use for the benchmark.- Returns:
- The result of the benchmark as a string.
-