Spaces:
Runtime error
Runtime error
switched to filepath audio
Browse files
app.py
CHANGED
|
@@ -18,21 +18,25 @@ seamless_client = Client("facebook/seamless_m4t")
|
|
| 18 |
|
| 19 |
|
| 20 |
|
| 21 |
-
def process_speech(
|
| 22 |
"""
|
| 23 |
processing sound using seamless_m4t
|
| 24 |
"""
|
| 25 |
-
print(f"audio : {
|
| 26 |
-
print(f"audio type : {type(
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
out = seamless_client.predict(
|
| 32 |
"S2TT",
|
| 33 |
"file",
|
| 34 |
None,
|
| 35 |
-
|
| 36 |
"",
|
| 37 |
"French",# source language
|
| 38 |
"English",# target language
|
|
@@ -306,7 +310,7 @@ with gr.Blocks(theme='ParityError/Anime') as iface :
|
|
| 306 |
image_output = gr.Markdown(label="output text")
|
| 307 |
image_button = gr.Button("process image")
|
| 308 |
with gr.Tab("speech to text translation"):
|
| 309 |
-
audio_input = gr.Audio(label="talk in french")
|
| 310 |
audio_output = gr.Markdown(label="output text")
|
| 311 |
audio_button = gr.Button("process audio")
|
| 312 |
text_button.click(process_and_query, inputs=text_input, outputs=text_output)
|
|
|
|
| 18 |
|
| 19 |
|
| 20 |
|
| 21 |
+
def process_speech(audio_input):
|
| 22 |
"""
|
| 23 |
processing sound using seamless_m4t
|
| 24 |
"""
|
| 25 |
+
print(f"audio : {audio_input}")
|
| 26 |
+
print(f"audio type : {type(audio_input)}")
|
| 27 |
+
|
| 28 |
+
try :
|
| 29 |
+
audio_name = f"{np.random.randint(0, 100)}.wav"
|
| 30 |
+
sr, data = audio_input
|
| 31 |
+
write(audio_name, sr, data.astype(np.int16))
|
| 32 |
+
audio_input = audio_name
|
| 33 |
+
except :
|
| 34 |
+
pass
|
| 35 |
out = seamless_client.predict(
|
| 36 |
"S2TT",
|
| 37 |
"file",
|
| 38 |
None,
|
| 39 |
+
audio_input, #audio_name
|
| 40 |
"",
|
| 41 |
"French",# source language
|
| 42 |
"English",# target language
|
|
|
|
| 310 |
image_output = gr.Markdown(label="output text")
|
| 311 |
image_button = gr.Button("process image")
|
| 312 |
with gr.Tab("speech to text translation"):
|
| 313 |
+
audio_input = gr.Audio(label="talk in french",type="filepath",sources="microphone")
|
| 314 |
audio_output = gr.Markdown(label="output text")
|
| 315 |
audio_button = gr.Button("process audio")
|
| 316 |
text_button.click(process_and_query, inputs=text_input, outputs=text_output)
|