Spaces:
Running
Running
First commit
Browse files- README.md +2 -2
- app.py +91 -0
- cmu_us_awb_arctic-wav-arctic_a0002.npy +3 -0
- cmu_us_bdl_arctic-wav-arctic_a0009.npy +3 -0
- cmu_us_clb_arctic-wav-arctic_a0144.npy +3 -0
- cmu_us_ksp_arctic-wav-arctic_b0087.npy +3 -0
- cmu_us_rms_arctic-wav-arctic_b0353.npy +3 -0
- cmu_us_slt_arctic-wav-arctic_a0508.npy +3 -0
- requirements.txt +11 -0
- speech.wav +0 -0
- style.css +49 -0
README.md
CHANGED
@@ -3,8 +3,8 @@ title: Text To Voice
|
|
3 |
emoji: π
|
4 |
colorFrom: green
|
5 |
colorTo: yellow
|
6 |
-
sdk:
|
7 |
-
sdk_version:
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
---
|
|
|
3 |
emoji: π
|
4 |
colorFrom: green
|
5 |
colorTo: yellow
|
6 |
+
sdk: streamlit
|
7 |
+
sdk_version: 1.28.2
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
---
|
app.py
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import numpy as np
|
3 |
+
import torch
|
4 |
+
from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5HifiGan
|
5 |
+
import soundfile as sf
|
6 |
+
from io import StringIO
|
7 |
+
|
8 |
+
# Load models outside of function calls for efficiency
|
9 |
+
@st.cache(allow_output_mutation=True)
|
10 |
+
def load_models():
|
11 |
+
model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts")
|
12 |
+
processor = SpeechT5Processor.from_pretrained("microsoft/speecht5_tts")
|
13 |
+
vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
|
14 |
+
return model, processor, vocoder
|
15 |
+
|
16 |
+
model, processor, vocoder = load_models()
|
17 |
+
|
18 |
+
# Load speaker embeddings
|
19 |
+
@st.cache(allow_output_mutation=True)
|
20 |
+
def get_speaker_embeddings():
|
21 |
+
speaker_embeddings = np.load("cmu_us_clb_arctic-wav-arctic_a0144.npy")
|
22 |
+
return torch.tensor(speaker_embeddings).unsqueeze(0)
|
23 |
+
|
24 |
+
speaker_embeddings = get_speaker_embeddings()
|
25 |
+
|
26 |
+
# Improved Styling (assuming style.css is present)
|
27 |
+
def local_css(file_name):
|
28 |
+
with open(file_name) as f:
|
29 |
+
st.markdown(f'<style>{f.read()}</style>', unsafe_allow_html=True)
|
30 |
+
|
31 |
+
local_css("style.css") # Apply custom CSS styles
|
32 |
+
|
33 |
+
# Streamlit Layout
|
34 |
+
st.title("Text-to-Voice Conversion")
|
35 |
+
st.markdown("Convert your text to speech using advanced AI models.")
|
36 |
+
|
37 |
+
# Function to convert text to speech
|
38 |
+
def text_to_speech(text):
|
39 |
+
try:
|
40 |
+
max_length = 100 # Set a max length as per model's capability
|
41 |
+
segments = [text[i:i+max_length] for i in range(0, len(text), max_length)]
|
42 |
+
audio_paths = []
|
43 |
+
|
44 |
+
for i, segment in enumerate(segments):
|
45 |
+
inputs = processor(text=segment, return_tensors="pt")
|
46 |
+
spectrogram = model.generate_speech(inputs["input_ids"], speaker_embeddings)
|
47 |
+
with torch.no_grad():
|
48 |
+
speech = vocoder(spectrogram)
|
49 |
+
audio_path = f"speech_segment_{i}.wav"
|
50 |
+
sf.write(audio_path, speech.numpy(), samplerate=16000)
|
51 |
+
audio_paths.append(audio_path)
|
52 |
+
|
53 |
+
return audio_paths
|
54 |
+
except Exception as e:
|
55 |
+
st.error(f"Error in text-to-speech conversion: {e}")
|
56 |
+
return []
|
57 |
+
|
58 |
+
# Function to combine audio segments
|
59 |
+
def combine_audio_segments(paths):
|
60 |
+
combined_speech = []
|
61 |
+
for path in paths:
|
62 |
+
data, samplerate = sf.read(path)
|
63 |
+
combined_speech.extend(data)
|
64 |
+
sf.write("combined_speech.wav", np.array(combined_speech), samplerate)
|
65 |
+
return "combined_speech.wav"
|
66 |
+
|
67 |
+
# Text Input and Conversion Button
|
68 |
+
text = st.text_area("Type your text here.")
|
69 |
+
|
70 |
+
if st.button("Convert"):
|
71 |
+
if text:
|
72 |
+
audio_paths = text_to_speech(text)
|
73 |
+
combined_audio_path = combine_audio_segments(audio_paths)
|
74 |
+
audio_bytes = open(combined_audio_path, 'rb').read()
|
75 |
+
st.audio(audio_bytes, format='audio/wav')
|
76 |
+
else:
|
77 |
+
st.error("Please enter some text to convert.")
|
78 |
+
|
79 |
+
# File Uploader and Conversion Button
|
80 |
+
uploaded_file = st.file_uploader("Upload a text file here", type=['txt'])
|
81 |
+
|
82 |
+
if uploaded_file is not None:
|
83 |
+
stringio = StringIO(uploaded_file.getvalue().decode("utf-8"))
|
84 |
+
text = stringio.read()
|
85 |
+
st.write(text)
|
86 |
+
|
87 |
+
if st.button("Convert Uploaded File", key="upload"):
|
88 |
+
audio_paths = text_to_speech(text)
|
89 |
+
combined_audio_path = combine_audio_segments(audio_paths)
|
90 |
+
audio_bytes = open(combined_audio_path, 'rb').read()
|
91 |
+
st.audio(audio_bytes, format='audio/wav')
|
cmu_us_awb_arctic-wav-arctic_a0002.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5db7a684ab490f21cec1628e00d461a184e369fe4eafb1ee441a796faf4ab6ae
|
3 |
+
size 2176
|
cmu_us_bdl_arctic-wav-arctic_a0009.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:215326eae3a428af8934c385fbe043b36c72849ca17d1d013adeb189e6bd6962
|
3 |
+
size 2176
|
cmu_us_clb_arctic-wav-arctic_a0144.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf67b36c47edfb1851466a1dff081b436bc6809b5ebc12811d9df0c0d0f28d0e
|
3 |
+
size 2176
|
cmu_us_ksp_arctic-wav-arctic_b0087.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f6c5c2a38c2e400179019c560a74c4322f4ee13beda22ee601807545edee283e
|
3 |
+
size 2176
|
cmu_us_rms_arctic-wav-arctic_b0353.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a49dac3e9c3a71a4dbca4c364233c7915ae6e0cb71b2ceaed97296231b95cb50
|
3 |
+
size 2176
|
cmu_us_slt_arctic-wav-arctic_a0508.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f71ffadda3f3a4de079740a0b34963824dc644d9d5442283bd0a2b0d4f44ff0b
|
3 |
+
size 2176
|
requirements.txt
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
streamlit_option_menu == 0.3.2
|
2 |
+
requests==2.28.1
|
3 |
+
streamlit-lottie==0.0.3
|
4 |
+
times==0.7
|
5 |
+
htbuilder==0.6.1
|
6 |
+
transformers==4.29.2
|
7 |
+
torch==2.0.1
|
8 |
+
soundfile==0.12.1
|
9 |
+
torchaudio == 2.0.2
|
10 |
+
sentencepiece==0.1.99
|
11 |
+
streamlit
|
speech.wav
ADDED
Binary file (10.3 kB). View file
|
|
style.css
ADDED
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
/* Custom CSS for Streamlit App */
|
2 |
+
|
3 |
+
/* Header Style */
|
4 |
+
header {
|
5 |
+
background-color: #5072A7; /* A soothing shade of blue */
|
6 |
+
color: white; /* White text for contrast */
|
7 |
+
padding: 10px; /* Adding padding for spacing */
|
8 |
+
}
|
9 |
+
|
10 |
+
/* Main Body Style */
|
11 |
+
body {
|
12 |
+
font-family: Arial, sans-serif; /* Standard readable font stack */
|
13 |
+
color: #333333; /* Dark grey color for main text */
|
14 |
+
line-height: 1.6; /* Improved line spacing for readability */
|
15 |
+
}
|
16 |
+
|
17 |
+
/* Titles and Headings */
|
18 |
+
h1, h2, h3 {
|
19 |
+
color: #333366; /* Darker shade of blue for headings */
|
20 |
+
margin-bottom: 10px; /* Adding margin for separation */
|
21 |
+
}
|
22 |
+
|
23 |
+
/* Buttons */
|
24 |
+
button {
|
25 |
+
background-color: #4CAF50; /* Green background for buttons */
|
26 |
+
color: white; /* White text for contrast */
|
27 |
+
padding: 10px 20px; /* Padding for button size */
|
28 |
+
margin: 10px 0; /* Margin for spacing */
|
29 |
+
border: none; /* Removing default border */
|
30 |
+
cursor: pointer; /* Pointer cursor for usability */
|
31 |
+
border-radius: 5px; /* Rounded corners */
|
32 |
+
transition: background-color 0.3s ease; /* Smooth hover transition */
|
33 |
+
}
|
34 |
+
|
35 |
+
button:hover {
|
36 |
+
background-color: #45a049; /* Darker green on hover */
|
37 |
+
}
|
38 |
+
|
39 |
+
/* Footer Style */
|
40 |
+
footer {
|
41 |
+
background-color: #333333; /* Dark background for footer */
|
42 |
+
color: white; /* White text for contrast */
|
43 |
+
text-align: center; /* Center-aligning text */
|
44 |
+
padding: 10px; /* Padding for spacing */
|
45 |
+
position: fixed; /* Fixed positioning */
|
46 |
+
left: 0;
|
47 |
+
bottom: 0;
|
48 |
+
width: 100%; /* Full width */
|
49 |
+
}
|