Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,57 +1,57 @@
|
|
1 |
-
import streamlit as st
|
2 |
-
import requests
|
3 |
-
|
4 |
-
API_URL_FASTSPEECH2 = "https://api-inference.huggingface.co/models/facebook/fastspeech2-en-ljspeech"
|
5 |
-
API_URL_DISTIL_WHISPER = "https://api-inference.huggingface.co/models/distil-whisper/distil-large-v3"
|
6 |
-
API_URL_WAV2VEC2 = "https://api-inference.huggingface.co/models/facebook/wav2vec2-large-960h"
|
7 |
-
|
8 |
-
def query(api_url, payload):
|
9 |
-
response = requests.post(api_url, json=payload)
|
10 |
-
if response.status_code == 200:
|
11 |
-
return response.content
|
12 |
-
else:
|
13 |
-
st.error(f"Erreur {response.status_code}: {response.text}")
|
14 |
-
return None
|
15 |
-
|
16 |
-
def query_audio(api_url, audio_file):
|
17 |
-
response = requests.post(api_url, files={"file": audio_file})
|
18 |
-
if response.status_code == 200:
|
19 |
-
return response.json()
|
20 |
-
else:
|
21 |
-
st.error(f"Erreur {response.status_code}: {response.text}")
|
22 |
-
return None
|
23 |
-
|
24 |
-
def main():
|
25 |
-
|
26 |
-
st.title("Text to speech App")
|
27 |
-
# Sidebar pour la navigation entre les modèles
|
28 |
-
st.sidebar.image("
|
29 |
-
model_selection = st.sidebar.selectbox("Sélectionnez le modèle", ["Text-to-Speech (fastspeech2-en-ljspeech)", "Speech-to-Text (distil-whisper/distil-large-v3)"])
|
30 |
-
|
31 |
-
if model_selection == "Text-to-Speech (fastspeech2-en-ljspeech)":
|
32 |
-
st.subheader("Text-to-Speech (fastspeech2-en-ljspeech)")
|
33 |
-
st.image("
|
34 |
-
st.text("le modele prend environ 20s pou se charger veiller patienter quelques secondes")
|
35 |
-
user_input_tts = st.text_area("Entrez le texte que vous souhaitez convertir en audio:")
|
36 |
-
|
37 |
-
if st.button("Convertir le texte en audio avec fastspeech2-en-ljspeech"):
|
38 |
-
payload_tts = {"inputs": user_input_tts}
|
39 |
-
audio_bytes = query(API_URL_FASTSPEECH2, payload_tts)
|
40 |
-
|
41 |
-
if audio_bytes:
|
42 |
-
st.audio(audio_bytes, format="audio/wav", start_time=0)
|
43 |
-
|
44 |
-
elif model_selection == "Speech-to-Text (distil-whisper/distil-large-v3)":
|
45 |
-
st.subheader("Speech-to-Text (distil-whisper/distil-large-v3)")
|
46 |
-
st.image("
|
47 |
-
audio_file = st.file_uploader("Téléchargez votre fichier audio", type=["wav", "mp3", "flac"])
|
48 |
-
|
49 |
-
if st.button("Convertir l'audio en texte avec distil-whisper/distil-large-v3") and audio_file is not None:
|
50 |
-
result_stt = query_audio(API_URL_DISTIL_WHISPER, audio_file)
|
51 |
-
|
52 |
-
if result_stt:
|
53 |
-
st.subheader("Texte transcrit:")
|
54 |
-
st.write(result_stt.get("text", "Aucune transcription trouvée."))
|
55 |
-
|
56 |
-
if __name__ == "__main__":
|
57 |
-
main()
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import requests
|
3 |
+
|
4 |
+
API_URL_FASTSPEECH2 = "https://api-inference.huggingface.co/models/facebook/fastspeech2-en-ljspeech"
|
5 |
+
API_URL_DISTIL_WHISPER = "https://api-inference.huggingface.co/models/distil-whisper/distil-large-v3"
|
6 |
+
API_URL_WAV2VEC2 = "https://api-inference.huggingface.co/models/facebook/wav2vec2-large-960h"
|
7 |
+
|
8 |
+
def query(api_url, payload):
|
9 |
+
response = requests.post(api_url, json=payload)
|
10 |
+
if response.status_code == 200:
|
11 |
+
return response.content
|
12 |
+
else:
|
13 |
+
st.error(f"Erreur {response.status_code}: {response.text}")
|
14 |
+
return None
|
15 |
+
|
16 |
+
def query_audio(api_url, audio_file):
|
17 |
+
response = requests.post(api_url, files={"file": audio_file})
|
18 |
+
if response.status_code == 200:
|
19 |
+
return response.json()
|
20 |
+
else:
|
21 |
+
st.error(f"Erreur {response.status_code}: {response.text}")
|
22 |
+
return None
|
23 |
+
|
24 |
+
def main():
|
25 |
+
|
26 |
+
st.title("Text to speech App")
|
27 |
+
# Sidebar pour la navigation entre les modèles
|
28 |
+
st.sidebar.image("istockphoto-1391947389-612x612.jpg")
|
29 |
+
model_selection = st.sidebar.selectbox("Sélectionnez le modèle", ["Text-to-Speech (fastspeech2-en-ljspeech)", "Speech-to-Text (distil-whisper/distil-large-v3)"])
|
30 |
+
|
31 |
+
if model_selection == "Text-to-Speech (fastspeech2-en-ljspeech)":
|
32 |
+
st.subheader("Text-to-Speech (fastspeech2-en-ljspeech)")
|
33 |
+
st.image("image2.png")
|
34 |
+
st.text("le modele prend environ 20s pou se charger veiller patienter quelques secondes")
|
35 |
+
user_input_tts = st.text_area("Entrez le texte que vous souhaitez convertir en audio:")
|
36 |
+
|
37 |
+
if st.button("Convertir le texte en audio avec fastspeech2-en-ljspeech"):
|
38 |
+
payload_tts = {"inputs": user_input_tts}
|
39 |
+
audio_bytes = query(API_URL_FASTSPEECH2, payload_tts)
|
40 |
+
|
41 |
+
if audio_bytes:
|
42 |
+
st.audio(audio_bytes, format="audio/wav", start_time=0)
|
43 |
+
|
44 |
+
elif model_selection == "Speech-to-Text (distil-whisper/distil-large-v3)":
|
45 |
+
st.subheader("Speech-to-Text (distil-whisper/distil-large-v3)")
|
46 |
+
st.image("istockphoto-1391947389-612x612.jpg")
|
47 |
+
audio_file = st.file_uploader("Téléchargez votre fichier audio", type=["wav", "mp3", "flac"])
|
48 |
+
|
49 |
+
if st.button("Convertir l'audio en texte avec distil-whisper/distil-large-v3") and audio_file is not None:
|
50 |
+
result_stt = query_audio(API_URL_DISTIL_WHISPER, audio_file)
|
51 |
+
|
52 |
+
if result_stt:
|
53 |
+
st.subheader("Texte transcrit:")
|
54 |
+
st.write(result_stt.get("text", "Aucune transcription trouvée."))
|
55 |
+
|
56 |
+
if __name__ == "__main__":
|
57 |
+
main()
|