Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import gradio as gr
|
3 |
+
from speechbrain.pretrained import Tacotron2, HIFIGAN
|
4 |
+
from scipy.io.wavfile import write
|
5 |
+
|
6 |
+
# Load TTS and vocoder models
|
7 |
+
tacotron2 = Tacotron2.from_hparams(source="speechbrain/tts-tacotron2-ljspeech", savedir="tmpdir_tts")
|
8 |
+
hifi_gan = HIFIGAN.from_hparams(source="speechbrain/tts-hifigan-ljspeech", savedir="tmpdir_vocoder")
|
9 |
+
|
10 |
+
# Function to generate speech from text
|
11 |
+
def text_to_speech(text):
|
12 |
+
mel_output, mel_length, alignment = tacotron2.encode_text(text)
|
13 |
+
waveforms, _, _ = hifi_gan.decode_batch(mel_output)
|
14 |
+
audio_path = "tts_output.wav"
|
15 |
+
write(audio_path, 22050, waveforms.squeeze(1).cpu().numpy())
|
16 |
+
return audio_path
|
17 |
+
|
18 |
+
# Gradio interface for TTS
|
19 |
+
def chatbot_response(text):
|
20 |
+
audio_path = text_to_speech(text)
|
21 |
+
return audio_path
|
22 |
+
|
23 |
+
gr_interface = gr.Interface(fn=chatbot_response, inputs="text", outputs="audio")
|
24 |
+
|
25 |
+
# Streamlit app setup
|
26 |
+
st.title("Text-to-Speech Chatbot")
|
27 |
+
st.write("This application converts text into speech using SpeechBrain.")
|
28 |
+
|
29 |
+
# Embed Gradio interface in Streamlit
|
30 |
+
with st.spinner("Loading Gradio interface..."):
|
31 |
+
gr_interface.launch(share=False, inbrowser=True)
|