public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-18 13:17:59 +02:00
Code Issues Packages Projects Releases Wiki Activity

whisper : remove speed_up and phase_vocoder* functions (#2198)

Browse Source 
* whisper : fix cast warning

* whisper : remove phase_vocoder functions, ref #2195

* whisper : remove speed_up from whisper_full_params, closes #2195
This commit is contained in:
Borislav Stanimirov
2024-05-31 11:37:29 +03:00
committed by GitHub
parent b87494bb8f
commit af5833e298
20 changed files with 14 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files
bindings/go/examples/go-whisper/flags.go
-9
View File
@@ -68,10 +68,6 @@ func (flags *Flags) GetOut() string {
return strings.ToLower(flags.Lookup("out").Value.String())
}
func (flags *Flags) IsSpeedup() bool {
return flags.Lookup("speedup").Value.String() == "true"
}
func (flags *Flags) IsTokens() bool {
return flags.Lookup("tokens").Value.String() == "true"
}
@@ -111,10 +107,6 @@ func (flags *Flags) SetParams(context whisper.Context) error {
fmt.Fprintf(flags.Output(), "Setting duration to %v\n", duration)
context.SetDuration(duration)
}
if flags.IsSpeedup() {
fmt.Fprintf(flags.Output(), "Setting speedup to true\n")
context.SetSpeedup(true)
}
if threads := flags.GetThreads(); threads != 0 {
fmt.Fprintf(flags.Output(), "Setting threads to %d\n", threads)
context.SetThreads(threads)
@@ -146,7 +138,6 @@ func registerFlags(flag *Flags) {
flag.Duration("offset", 0, "Time offset")
flag.Duration("duration", 0, "Duration of audio to process")
flag.Uint("threads", 0, "Number of threads to use")
flag.Bool("speedup", false, "Enable speedup")
flag.Uint("max-len", 0, "Maximum segment length in characters")
flag.Uint("max-tokens", 0, "Maximum tokens per segment")
flag.Float64("word-thold", 0, "Maximum segment score")
bindings/go/params.go
-7
View File
@@ -47,10 +47,6 @@ func (p *Params) SetPrintTimestamps(v bool) {
p.print_timestamps = toBool(v)
}
func (p *Params) SetSpeedup(v bool) {
p.speed_up = toBool(v)
}
// Set language id
func (p *Params) SetLanguage(lang int) error {
if lang == -1 {
@@ -177,9 +173,6 @@ func (p *Params) String() string {
if p.token_timestamps {
str += " token_timestamps"
}
if p.speed_up {
str += " speed_up"
}
return str + ">"
}
bindings/go/pkg/whisper/context.go
-5
View File
@@ -76,11 +76,6 @@ func (context *context) SetTranslate(v bool) {
context.params.SetTranslate(v)
}
// Set speedup flag
func (context *context) SetSpeedup(v bool) {
context.params.SetSpeedup(v)
}
func (context *context) SetSplitOnWord(v bool) {
context.params.SetSplitOnWord(v)
}
bindings/go/pkg/whisper/interface.go
-1
View File
@@ -41,7 +41,6 @@ type Context interface {
SetOffset(time.Duration) // Set offset
SetDuration(time.Duration) // Set duration
SetThreads(uint) // Set number of threads to use
SetSpeedup(bool) // Set speedup flag
SetSplitOnWord(bool) // Set split on word flag
SetTokenThreshold(float32) // Set timestamp token probability threshold
SetTokenSumThreshold(float32) // Set timestamp token sum probability threshold
bindings/java/src/main/java/io/github/ggerganov/whispercpp/WhisperCppJnaLibrary.java
+1 -9
View File
@@ -20,7 +20,7 @@ public interface WhisperCppJnaLibrary extends Library {
* @return Whisper context on success, null on failure
*/
Pointer whisper_init_from_file(String path_model);
/**
* Provides default params which can be used with `whisper_init_from_file_with_params()` etc.
* Because this function allocates memory for the params, the caller must call either:
@@ -304,14 +304,6 @@ public interface WhisperCppJnaLibrary extends Library {
/** Language id associated with the provided state */
int whisper_full_lang_id_from_state(Pointer state);
/**
* 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.
* @return 0 on success
*/
int whisper_pcm_to_mel_phase_vocoder(Pointer ctx, final float[] samples, int n_samples, int n_threads);
int whisper_pcm_to_mel_phase_vocoder_with_state(Pointer ctx, Pointer state, final float[] samples, int n_samples, int n_threads);
/** Get the start time of the specified segment. */
long whisper_full_get_segment_t0(Pointer ctx, int i_segment);
bindings/java/src/main/java/io/github/ggerganov/whispercpp/params/WhisperFullParams.java
+1 -9
View File
@@ -129,14 +129,6 @@ public class WhisperFullParams extends Structure {
/** Maximum tokens per segment (0, default = no limit) */
public int max_tokens;
/** Flag to speed up the audio by 2x using Phase Vocoder. (default = false) */
public CBool speed_up;
/** Flag to speed up the audio by 2x using Phase Vocoder. (default = false) */
public void speedUp(boolean enable) {
speed_up = enable ? CBool.TRUE : CBool.FALSE;
}
/** Overwrite the audio context size (0 = use default). */
public int audio_ctx;
@@ -321,7 +313,7 @@ public class WhisperFullParams extends Structure {
return Arrays.asList("strategy", "n_threads", "n_max_text_ctx", "offset_ms", "duration_ms", "translate",
"no_context", "single_segment", "no_timestamps",
"print_special", "print_progress", "print_realtime", "print_timestamps", "token_timestamps",
"thold_pt", "thold_ptsum", "max_len", "split_on_word", "max_tokens", "speed_up", "audio_ctx",
"thold_pt", "thold_ptsum", "max_len", "split_on_word", "max_tokens", "audio_ctx",
"tdrz_enable", "suppress_regex", "initial_prompt", "prompt_tokens", "prompt_n_tokens", "language", "detect_language",
"suppress_blank", "suppress_non_speech_tokens", "temperature", "max_initial_ts", "length_penalty",
"temperature_inc", "entropy_thold", "logprob_thold", "no_speech_thold", "greedy", "beam_search",
bindings/ruby/ext/ruby_whisper.cpp
-8
View File
@@ -311,12 +311,6 @@ static VALUE ruby_whisper_params_get_split_on_word(VALUE self) {
static VALUE ruby_whisper_params_set_split_on_word(VALUE self, VALUE value) {
BOOL_PARAMS_SETTER(self, split_on_word, value)
}
static VALUE ruby_whisper_params_get_speed_up(VALUE self) {
BOOL_PARAMS_GETTER(self, speed_up)
}
static VALUE ruby_whisper_params_set_speed_up(VALUE self, VALUE value) {
BOOL_PARAMS_SETTER(self, speed_up, value)
}
static VALUE ruby_whisper_params_get_diarize(VALUE self) {
ruby_whisper_params *rwp;
Data_Get_Struct(self, ruby_whisper_params, rwp);
@@ -408,8 +402,6 @@ void Init_whisper() {
rb_define_method(cParams, "token_timestamps=", ruby_whisper_params_set_token_timestamps, 1);
rb_define_method(cParams, "split_on_word", ruby_whisper_params_get_split_on_word, 0);
rb_define_method(cParams, "split_on_word=", ruby_whisper_params_set_split_on_word, 1);
rb_define_method(cParams, "speed_up", ruby_whisper_params_get_speed_up, 0);
rb_define_method(cParams, "speed_up=", ruby_whisper_params_set_speed_up, 1);
rb_define_method(cParams, "diarize", ruby_whisper_params_get_diarize, 0);
rb_define_method(cParams, "diarize=", ruby_whisper_params_set_diarize, 1);
bindings/ruby/tests/test_whisper.rb
-7
View File
@@ -117,13 +117,6 @@ class TestWhisper < Test::Unit::TestCase
assert !@params.split_on_word
end
def test_speed_up
@params.speed_up = true
assert @params.speed_up
@params.speed_up = false
assert !@params.speed_up
end
def test_whisper
@whisper = Whisper::Context.new(File.join(TOPDIR, '..', '..', 'models', 'ggml-base.en.bin'))
params = Whisper::Params.new