Spaces:
Running
Running
Update util.py
Browse files
util.py
CHANGED
@@ -46,8 +46,8 @@ def get_subtitles(video_url):
|
|
46 |
from youtube_transcript_api import YouTubeTranscriptApi
|
47 |
|
48 |
|
49 |
-
def restore_punctuation(text):
|
50 |
-
model =
|
51 |
result = model.restore_punctuation(text)
|
52 |
return result
|
53 |
|
@@ -124,19 +124,3 @@ def display(text):
|
|
124 |
return formatted_sentences
|
125 |
|
126 |
|
127 |
-
|
128 |
-
def pipeline(url, model, tokenizer):
|
129 |
-
trans, sub = get_subtitles(url)
|
130 |
-
sub = restore_punctuation(sub)
|
131 |
-
vie_sub = translate_long(sub)
|
132 |
-
vie_sub = processed(vie_sub)
|
133 |
-
chunks = split_into_chunks(vie_sub, 700, 2)
|
134 |
-
sum_para = []
|
135 |
-
for i in chunks:
|
136 |
-
tmp = summarize(i, model, tokenizer, num_beams=3)
|
137 |
-
sum_para.append(tmp)
|
138 |
-
suma = ''.join(sum_para)
|
139 |
-
del sub, vie_sub, sum_para, chunks
|
140 |
-
suma = post_processing(suma)
|
141 |
-
re = display(suma)
|
142 |
-
return re
|
|
|
46 |
from youtube_transcript_api import YouTubeTranscriptApi
|
47 |
|
48 |
|
49 |
+
def restore_punctuation(text, model_restore):
|
50 |
+
model = model_restore
|
51 |
result = model.restore_punctuation(text)
|
52 |
return result
|
53 |
|
|
|
124 |
return formatted_sentences
|
125 |
|
126 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|