Spaces:
Sleeping
Sleeping
Commit
·
875c690
1
Parent(s):
35839aa
deciding on final language
Browse files
app.py
CHANGED
@@ -10,18 +10,13 @@ model = whisper.load_model("small")
|
|
10 |
|
11 |
# A table to look up all the languages
|
12 |
language_id_lookup = {
|
|
|
13 |
"English" : "en",
|
|
|
14 |
"German" : "de",
|
15 |
-
"Greek" : "el",
|
16 |
"Spanish" : "es",
|
17 |
-
"Finnish" : "fi",
|
18 |
"Russian" : "ru",
|
19 |
-
"Hungarian" : "hu",
|
20 |
-
"Dutch" : "nl",
|
21 |
"French" : "fr",
|
22 |
-
'Polish' : "pl",
|
23 |
-
'Portuguese': "pt",
|
24 |
-
'Italian' : "it",
|
25 |
}
|
26 |
|
27 |
|
@@ -69,26 +64,22 @@ def predict(audio, language, mic_audio=None):
|
|
69 |
title = "Demo for Whisper -> Something -> XLS-R"
|
70 |
|
71 |
description = """
|
72 |
-
<b>How to use:</b> Upload an audio file or record using the microphone. The audio is
|
73 |
-
|
|
|
74 |
"""
|
75 |
|
76 |
gr.Interface(
|
77 |
fn=predict,
|
78 |
inputs=[
|
79 |
gr.Audio(label="Upload Speech", source="upload", type="filepath"),
|
80 |
-
gr.inputs.Dropdown(['
|
|
|
|
|
81 |
'German Text',
|
82 |
-
'Greek Text',
|
83 |
'Spanish Text',
|
84 |
-
'Finnish Text',
|
85 |
'Russian Text',
|
86 |
-
'Hungarian Text',
|
87 |
-
'Dutch Text',
|
88 |
'French Text',
|
89 |
-
'Polish Text',
|
90 |
-
'Portuguese Text',
|
91 |
-
'Italian Text',
|
92 |
'Detect Language'], type="value", default='English Text', label="Select the Language of the that you are speaking in."),
|
93 |
gr.Audio(label="Record Speech", source="microphone", type="filepath"),
|
94 |
],
|
|
|
10 |
|
11 |
# A table to look up all the languages
|
12 |
language_id_lookup = {
|
13 |
+
"Arabic" : "ar",
|
14 |
"English" : "en",
|
15 |
+
"Chinese" : "zh",
|
16 |
"German" : "de",
|
|
|
17 |
"Spanish" : "es",
|
|
|
18 |
"Russian" : "ru",
|
|
|
|
|
19 |
"French" : "fr",
|
|
|
|
|
|
|
20 |
}
|
21 |
|
22 |
|
|
|
64 |
title = "Demo for Whisper -> Something -> XLS-R"
|
65 |
|
66 |
description = """
|
67 |
+
<b>How to use:</b> Upload an audio file or record using the microphone. The audio is into the whisper model developed by openai.
|
68 |
+
The output is the text transcription of the audio in the language you inputted. If you asked the model to detect a language, it will
|
69 |
+
tell you what language it detected.
|
70 |
"""
|
71 |
|
72 |
gr.Interface(
|
73 |
fn=predict,
|
74 |
inputs=[
|
75 |
gr.Audio(label="Upload Speech", source="upload", type="filepath"),
|
76 |
+
gr.inputs.Dropdown(['Arabic Text',
|
77 |
+
'Chinese Text',
|
78 |
+
'English Text',
|
79 |
'German Text',
|
|
|
80 |
'Spanish Text',
|
|
|
81 |
'Russian Text',
|
|
|
|
|
82 |
'French Text',
|
|
|
|
|
|
|
83 |
'Detect Language'], type="value", default='English Text', label="Select the Language of the that you are speaking in."),
|
84 |
gr.Audio(label="Record Speech", source="microphone", type="filepath"),
|
85 |
],
|