Spaces:
Sleeping
Sleeping
Commit
·
e77e1d8
1
Parent(s):
3146b8d
Update app.py
Browse files
app.py
CHANGED
@@ -7,18 +7,13 @@ import whisperx
|
|
7 |
|
8 |
model = WhisperModel('large-v2', device="cuda", compute_type="float16")
|
9 |
|
10 |
-
def speech_to_text(mic=None, file=None, lang=None):
|
11 |
if mic is not None:
|
12 |
audio = mic
|
13 |
elif file is not None:
|
14 |
audio = file
|
15 |
else:
|
16 |
raise gr.Error("You must either provide a mic recording or a file")
|
17 |
-
|
18 |
-
if lang is None or lang == '':
|
19 |
-
lang = None
|
20 |
-
|
21 |
-
task = 'transcribe' if lang is None else 'translate'
|
22 |
print(lang, task)
|
23 |
|
24 |
time_start = time.time()
|
@@ -73,11 +68,11 @@ with gr.Blocks(title='Whisper Demo', theme=theme) as demo:
|
|
73 |
translate_btn = gr.Button("Translate audio")
|
74 |
trans_df = gr.DataFrame(label="Transcription dataframe", row_count=(0, "dynamic"), max_rows = 10, wrap=True, overflow_row_behaviour='paginate')
|
75 |
sys_info = gr.Markdown("")
|
76 |
-
transcribe_btn.click(speech_to_text,
|
77 |
[audio_in, file_in, drop_down],
|
78 |
[trans_df, sys_info]
|
79 |
)
|
80 |
-
translate_btn.click(speech_to_text,
|
81 |
[audio_in, file_in, drop_down],
|
82 |
[trans_df, sys_info])
|
83 |
|
|
|
7 |
|
8 |
model = WhisperModel('large-v2', device="cuda", compute_type="float16")
|
9 |
|
10 |
+
def speech_to_text(mic=None, file=None, lang=None, task='transcribe'):
|
11 |
if mic is not None:
|
12 |
audio = mic
|
13 |
elif file is not None:
|
14 |
audio = file
|
15 |
else:
|
16 |
raise gr.Error("You must either provide a mic recording or a file")
|
|
|
|
|
|
|
|
|
|
|
17 |
print(lang, task)
|
18 |
|
19 |
time_start = time.time()
|
|
|
68 |
translate_btn = gr.Button("Translate audio")
|
69 |
trans_df = gr.DataFrame(label="Transcription dataframe", row_count=(0, "dynamic"), max_rows = 10, wrap=True, overflow_row_behaviour='paginate')
|
70 |
sys_info = gr.Markdown("")
|
71 |
+
transcribe_btn.click(lambda x, y: speech_to_text(x, y, None, 'transcribe'),
|
72 |
[audio_in, file_in, drop_down],
|
73 |
[trans_df, sys_info]
|
74 |
)
|
75 |
+
translate_btn.click(lambda x, y, z: speech_to_text(x, y, None, 'translate'),
|
76 |
[audio_in, file_in, drop_down],
|
77 |
[trans_df, sys_info])
|
78 |
|