ed
/
whispercpp.py
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Configure language from python

This commit is contained in:
John Ryan 2023-08-14 13:56:36 +01:00
parent 7af678159c
commit d448db6dc3
1 changed files with 8 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
whispercpp.pyx
View File

@ -16,7 +16,6 @@ cimport numpy as cnp
cdef int SAMPLE_RATE = 16000 cdef int SAMPLE_RATE = 16000
cdef char* TEST_FILE = 'test.wav' cdef char* TEST_FILE = 'test.wav'
cdef char* DEFAULT_MODEL = 'tiny' cdef char* DEFAULT_MODEL = 'tiny'
cdef char* LANGUAGE = b'en'
cdef int N_THREADS = os.cpu_count() cdef int N_THREADS = os.cpu_count()
MODELS = { MODELS = {
@ -68,14 +67,14 @@ cdef cnp.ndarray[cnp.float32_t, ndim=1, mode="c"] load_audio(bytes file, int sr
return frames return frames
cdef whisper_full_params default_params() nogil: cdef whisper_full_params default_params(char* language) nogil:
cdef whisper_full_params params = whisper_full_default_params( cdef whisper_full_params params = whisper_full_default_params(
whisper_sampling_strategy.WHISPER_SAMPLING_GREEDY whisper_sampling_strategy.WHISPER_SAMPLING_GREEDY
) )
params.print_realtime = True params.print_realtime = True
params.print_progress = True params.print_progress = True
params.translate = False params.translate = False
params.language = <const char *> LANGUAGE params.language = <const char *>language
n_threads = N_THREADS n_threads = N_THREADS
return params return params
@ -96,14 +95,13 @@ cdef class Whisper:
else: else:
self.ctx = whisper_init_from_file(model_b) self.ctx = whisper_init_from_file(model_b)
self.params = default_params()
whisper_print_system_info() whisper_print_system_info()
def __dealloc__(self): def __dealloc__(self):
whisper_free(self.ctx) whisper_free(self.ctx)
def transcribe(self, filename=TEST_FILE): def transcribe(self, filename=TEST_FILE, lang="auto"):
print("Loading data..") print("Loading data..")
if (type(filename) == np.ndarray) : if (type(filename) == np.ndarray) :
temp = filename temp = filename
@ -115,9 +113,13 @@ cdef class Whisper:
cdef cnp.ndarray[cnp.float32_t, ndim=1, mode="c"] frames = temp cdef cnp.ndarray[cnp.float32_t, ndim=1, mode="c"] frames = temp
cdef char* c_string = lang # Convert bytes to char*
params = default_params(c_string)
print("Transcribing..") print("Transcribing..")
return whisper_full(self.ctx, self.params, &frames[0], len(frames))
return whisper_full(self.ctx, params, &frames[0], len(frames))
def extract_text(self, int res): def extract_text(self, int res):
print("Extracting text...") print("Extracting text...")