kobim/whisper-local-transcribe
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update transcribe.py

Browse Source 
better time keeping
This commit is contained in:
Kristofer Rolf Söderström
2023-04-20 20:13:54 +02:00
committed by GitHub
parent dfe967bd58
commit 4e1c709f43
1 changed files with 10 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
transcribe.py
View File

@@ -1,5 +1,7 @@
import whisper import whisper
import glob, os import glob, os
#import torch #uncomment if using torch with cuda, below too
import datetime
def transcribe(path, file_type, model=None, language=None, verbose=False): def transcribe(path, file_type, model=None, language=None, verbose=False):
'''Implementation of OpenAI's whisper model. Downloads model, transcribes audio files in a folder and returns the text files with transcriptions''' '''Implementation of OpenAI's whisper model. Downloads model, transcribes audio files in a folder and returns the text files with transcriptions'''
@@ -11,6 +13,11 @@ def transcribe(path, file_type, model=None, language=None, verbose=False):
glob_file = glob.glob(path+'/*{}'.format(file_type)) glob_file = glob.glob(path+'/*{}'.format(file_type))
#if torch.cuda.is_available():
# generator = torch.Generator('cuda').manual_seed(42)
#else:
# generator = torch.Generator().manual_seed(42)
print('Using {} model'.format(model)) print('Using {} model'.format(model))
print('File type is {}'.format(file_type)) print('File type is {}'.format(file_type))
print('Language is being detected automatically for each file') print('Language is being detected automatically for each file')
@@ -34,15 +41,15 @@ def transcribe(path, file_type, model=None, language=None, verbose=False):
end=[] end=[]
text=[] text=[]
for i in range(len(result['segments'])): for i in range(len(result['segments'])):
start.append(result['segments'][i]['start']) start.append(str(datetime.timedelta(seconds=(result['segments'][i]['start']))))
end.append(result['segments'][i]['end']) end.append(str(datetime.timedelta(seconds=(result['segments'][i]['end']))))
text.append(result['segments'][i]['text']) text.append(result['segments'][i]['text'])
with open("{}/transcriptions/{}.txt".format(path,title), 'w', encoding='utf-8') as file: with open("{}/transcriptions/{}.txt".format(path,title), 'w', encoding='utf-8') as file:
file.write(title) file.write(title)
file.write('\nIn seconds:') file.write('\nIn seconds:')
for i in range(len(result['segments'])): for i in range(len(result['segments'])):
file.writelines('\n[{:.2f} --> {:.2f}]:{}'.format(start[i], end[i], text[i])) file.writelines('\n[{} --> {}]:{}'.format(start[i], end[i], text[i]))
print('\nFinished file number {}.\n\n\n'.format(idx+1)) print('\nFinished file number {}.\n\n\n'.format(idx+1))