DHEIVER's picture
Update app.py
6db90a4 verified
import gradio as gr
import torch
import torchaudio
import numpy as np
from transformers import AutoProcessor, SeamlessM4Tv2Model
class TranslationModel:
def __init__(self):
self.model_name = "facebook/seamless-m4t-v2-large"
print("Loading model...")
self.processor = AutoProcessor.from_pretrained(self.model_name)
self.model = SeamlessM4Tv2Model.from_pretrained(self.model_name)
self.sample_rate = self.model.config.sampling_rate
self.languages = {
"English": "eng",
"Spanish": "spa",
"French": "fra",
"German": "deu",
"Italian": "ita",
"Portuguese": "por",
"Russian": "rus",
"Chinese": "cmn",
"Japanese": "jpn",
"Korean": "kor"
}
def translate_text(self, text, src_lang, tgt_lang, progress=gr.Progress()):
try:
progress(0.3, desc="Processing...")
inputs = self.processor(text=text, src_lang=self.languages[src_lang], return_tensors="pt")
progress(0.6, desc="Generating...")
audio_array = self.model.generate(**inputs, tgt_lang=self.languages[tgt_lang])[0].cpu().numpy().squeeze()
progress(1.0, desc="Complete")
return (self.sample_rate, audio_array)
except Exception as e:
raise gr.Error(str(e))
def translate_audio(self, audio_path, tgt_lang, progress=gr.Progress()):
if not audio_path:
raise gr.Error("Please upload an audio file")
try:
progress(0.3, desc="Processing...")
audio, orig_freq = torchaudio.load(audio_path)
audio = torchaudio.functional.resample(audio, orig_freq=orig_freq, new_freq=16000)
progress(0.6, desc="Translating...")
inputs = self.processor(audios=audio, return_tensors="pt")
audio_array = self.model.generate(**inputs, tgt_lang=self.languages[tgt_lang])[0].cpu().numpy().squeeze()
progress(1.0, desc="Complete")
return (self.sample_rate, audio_array)
except Exception as e:
raise gr.Error(str(e))
css = """
:root {
--primary-color: #2D3648;
--secondary-color: #5E6AD2;
--background-color: #F5F7FF;
--text-color: #2D3648;
--border-radius: 12px;
--spacing: 20px;
}
.gradio-container {
background-color: var(--background-color) !important;
}
.main-container {
max-width: 1200px !important;
margin: 0 auto !important;
padding: var(--spacing) !important;
}
.app-header {
text-align: center;
padding: 40px 20px;
background: linear-gradient(45deg, var(--primary-color), var(--secondary-color));
border-radius: var(--border-radius);
color: white !important;
margin-bottom: var(--spacing);
}
.app-title {
font-size: 2.5em;
font-weight: 700;
margin-bottom: 10px;
color: white !important;
}
.app-subtitle {
font-size: 1.2em;
opacity: 0.9;
color: white !important;
}
.content-block {
background: white;
padding: var(--spacing);
border-radius: var(--border-radius);
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.05);
margin-bottom: var(--spacing);
}
.gr-button {
background: var(--secondary-color) !important;
border: none !important;
color: white !important;
}
.gr-button:hover {
box-shadow: 0 4px 10px rgba(94, 106, 210, 0.3) !important;
transform: translateY(-1px);
}
.gr-input, .gr-select {
border-radius: 8px !important;
border: 2px solid #E5E7EB !important;
padding: 12px !important;
}
.gr-input:focus, .gr-select:focus {
border-color: var(--secondary-color) !important;
box-shadow: 0 0 0 3px rgba(94, 106, 210, 0.1) !important;
}
.gr-form {
background: white !important;
padding: var(--spacing) !important;
border-radius: var(--border-radius) !important;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.05) !important;
}
.gr-box {
border-radius: var(--border-radius) !important;
border: none !important;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.05) !important;
}
.footer {
text-align: center;
color: var(--text-color);
padding: var(--spacing);
opacity: 0.8;
}
/* Custom Tabs Styling */
.tab-nav {
background: white !important;
padding: 10px !important;
border-radius: var(--border-radius) !important;
margin-bottom: var(--spacing) !important;
}
.tab-nav button {
border-radius: 8px !important;
padding: 12px 24px !important;
}
.tab-nav button.selected {
background: var(--secondary-color) !important;
color: white !important;
}
"""
def create_ui():
model = TranslationModel()
with gr.Blocks(css=css, title="AI Language Translator") as demo:
gr.HTML(
"""
<div class="app-header">
<div class="app-title">AI Language Translator</div>
<div class="app-subtitle">Powered by Neural Machine Translation</div>
</div>
"""
)
with gr.Tabs():
# Text Translation Tab
with gr.Tab("Text to Speech"):
with gr.Column(variant="panel"):
gr.Markdown("### Enter Text")
text_input = gr.Textbox(
label="",
placeholder="Type or paste your text here...",
lines=4
)
with gr.Row():
src_lang = gr.Dropdown(
choices=sorted(model.languages.keys()),
value="English",
label="From"
)
tgt_lang = gr.Dropdown(
choices=sorted(model.languages.keys()),
value="Spanish",
label="To"
)
translate_btn = gr.Button("Translate", size="lg")
gr.Markdown("### Translation Output")
audio_output = gr.Audio(
label="",
type="numpy",
show_download_button=True
)
# Audio Translation Tab
with gr.Tab("Speech to Speech"):
with gr.Column(variant="panel"):
gr.Markdown("### Upload Audio")
audio_input = gr.Audio(
label="",
type="filepath",
sources=["upload", "microphone"]
)
tgt_lang_audio = gr.Dropdown(
choices=sorted(model.languages.keys()),
value="English",
label="Translate to"
)
translate_audio_btn = gr.Button("Translate Audio", size="lg")
gr.Markdown("### Translation Output")
audio_output_from_audio = gr.Audio(
label="",
type="numpy",
show_download_button=True
)
gr.HTML(
"""
<div class="footer">
Built with ❤️ using Meta's SeamlessM4T and Gradio
</div>
"""
)
# Event handlers
translate_btn.click(
fn=model.translate_text,
inputs=[text_input, src_lang, tgt_lang],
outputs=audio_output
)
translate_audio_btn.click(
fn=model.translate_audio,
inputs=[audio_input, tgt_lang_audio],
outputs=audio_output_from_audio
)
return demo
if __name__ == "__main__":
demo = create_ui()
demo.queue()
demo.launch()