Spaces:
Sleeping
Sleeping
skriller18
commited on
Commit
โข
ae68901
1
Parent(s):
1a0027d
Used API backend
Browse files- app.py +32 -13
- output.wav +0 -0
app.py
CHANGED
@@ -1,21 +1,18 @@
|
|
1 |
import torch
|
2 |
import torchaudio
|
3 |
-
from transformers import pipeline
|
4 |
import streamlit as st
|
|
|
|
|
|
|
|
|
5 |
|
6 |
model_id = '11mlabs/indri-0.1-124m-tts'
|
7 |
task = 'indri-tts'
|
8 |
|
9 |
-
pipe = pipeline(
|
10 |
-
task,
|
11 |
-
model=model_id,
|
12 |
-
#device=torch.device('cuda:0'), # Update this based on your hardware,
|
13 |
-
trust_remote_code=True
|
14 |
-
)
|
15 |
-
|
16 |
st.title("Indri")
|
17 |
st.subheader("Ultrafast multi-modal AI")
|
18 |
|
|
|
19 |
|
20 |
speakers = {
|
21 |
"[spkr_63]" : "๐ฌ๐ง ๐จ book reader",
|
@@ -33,7 +30,6 @@ speakers = {
|
|
33 |
"[spkr_66]" : "๐ฎ๐ณ ๐จ politician"
|
34 |
}
|
35 |
|
36 |
-
# Create a container for the speaker selection and text input
|
37 |
with st.container():
|
38 |
st.markdown("### Speaker Selection")
|
39 |
speaker_id = st.selectbox("Select a speaker:", options=list(speakers.keys()), format_func=lambda x: speakers[x])
|
@@ -41,10 +37,33 @@ with st.container():
|
|
41 |
st.markdown("### Text Input")
|
42 |
text_input = st.text_area("Enter text for TTS (max 200 characters):", max_chars=200)
|
43 |
|
44 |
-
|
45 |
if text_input:
|
46 |
-
|
47 |
-
|
48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
49 |
else:
|
50 |
st.warning("Please enter text to generate audio.")
|
|
|
|
|
|
|
|
1 |
import torch
|
2 |
import torchaudio
|
|
|
3 |
import streamlit as st
|
4 |
+
import requests
|
5 |
+
import streamlit.components.v1 as components
|
6 |
+
import asyncio
|
7 |
+
import io
|
8 |
|
9 |
model_id = '11mlabs/indri-0.1-124m-tts'
|
10 |
task = 'indri-tts'
|
11 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
st.title("Indri")
|
13 |
st.subheader("Ultrafast multi-modal AI")
|
14 |
|
15 |
+
baseUrl = "https://a8a9e62422722.notebooks.jarvislabs.net"
|
16 |
|
17 |
speakers = {
|
18 |
"[spkr_63]" : "๐ฌ๐ง ๐จ book reader",
|
|
|
30 |
"[spkr_66]" : "๐ฎ๐ณ ๐จ politician"
|
31 |
}
|
32 |
|
|
|
33 |
with st.container():
|
34 |
st.markdown("### Speaker Selection")
|
35 |
speaker_id = st.selectbox("Select a speaker:", options=list(speakers.keys()), format_func=lambda x: speakers[x])
|
|
|
37 |
st.markdown("### Text Input")
|
38 |
text_input = st.text_area("Enter text for TTS (max 200 characters):", max_chars=200)
|
39 |
|
40 |
+
async def generate_audio():
|
41 |
if text_input:
|
42 |
+
speaker_name = speakers[speaker_id]
|
43 |
+
|
44 |
+
response = requests.post(
|
45 |
+
f"{baseUrl}/tts",
|
46 |
+
json={
|
47 |
+
"text": text_input,
|
48 |
+
"speaker": speaker_name
|
49 |
+
},
|
50 |
+
headers={
|
51 |
+
"accept": "application/json",
|
52 |
+
"Content-Type": "application/json"
|
53 |
+
}
|
54 |
+
)
|
55 |
+
|
56 |
+
if response.ok:
|
57 |
+
audio_blob = response.content
|
58 |
+
|
59 |
+
audio_tensor, sample_rate = torchaudio.load(io.BytesIO(audio_blob))
|
60 |
+
|
61 |
+
torchaudio.save('output.wav', audio_tensor, sample_rate=sample_rate)
|
62 |
+
st.audio('output.wav')
|
63 |
+
else:
|
64 |
+
st.warning(f"Received invalid response format. Status Code: {response.status_code}, Response: {response.text}")
|
65 |
else:
|
66 |
st.warning("Please enter text to generate audio.")
|
67 |
+
|
68 |
+
if st.button("Generate Audio", key="generate_audio"):
|
69 |
+
asyncio.run(generate_audio())
|
output.wav
ADDED
Binary file (119 kB). View file
|
|