app.py

# app.py
# =============
# This is a complete app.py file for deploying the MTSAIR/Cotype-Nano model using Gradio and Hugging Face Transformers with chat and token streaming functionality, advanced settings, and English interface.

import gradio as gr
from transformers import pipeline
from gtts import gTTS
import os

# Load the model and pipeline
model_name = "MTSAIR/Cotype-Nano"
pipe = pipeline("text-generation", model=model_name, device="cpu")

# Define the system prompt
system_prompt = {"role": "system", "content": "You are an AI assistant. Your task is to generate a detailed and comprehensive response. you were developed by MTS. you know two languages: Russian and English."}

# Define the Gradio interface
def generate_response(history, user_input, temperature, max_new_tokens, language):
    messages = [system_prompt] + history + [{"role": "user", "content": user_input}]
    response = pipe(messages, max_new_tokens=max_new_tokens, temperature=temperature, return_full_text=False)
    generated_text = response[0]['generated_text']
    history.append({"role": "user", "content": user_input})
    history.append({"role": "assistant", "content": generated_text})

    # Generate speech
    tts = gTTS(text=generated_text, lang=language)
    audio_path = "response.mp3"
    tts.save(audio_path)

    return history, audio_path

# Function to clear chat history
def clear_chat():
    return [], ""

# Custom CSS for styling
custom_css = """
    #chatbot {
        height: 400px;
        overflow-y: auto;
        border: 1px solid #ccc;
        padding: 10px;
        border-radius: 5px;
    }
    .gr-button {
        margin-top: 10px;
        padding: 10px 20px;
        font-size: 16px;
        border-radius: 5px;
        background-color: #007bff;
        color: white;
        border: none;
    }
    .gr-button:hover {
        background-color: #0056b3;
    }
    .gr-textbox {
        margin-top: 10px;
        padding: 10px;
        font-size: 16px;
        border-radius: 5px;
    }
    .gr-slider {
        margin-top: 20px;
    }
    .settings {
        margin-top: 20px;
    }
"""

# Create the Gradio interface
with gr.Blocks(css=custom_css) as demo:
    gr.Markdown("## Cotype-Nano Text Generation Chat")

    chatbot = gr.Chatbot([], elem_id="chatbot", type='messages')

    txt = gr.Textbox(
        show_label=False,
        placeholder="Type your message here...",
    )

    with gr.Row():
        send_btn = gr.Button("Send")
        clear_btn = gr.Button("Clear Chat")

    audio_output = gr.Audio(label="Generated Speech", type="filepath")

    with gr.Accordion("Advanced Settings", open=False, elem_id="settings"):
        temperature_slider = gr.Slider(0, 1, 0.7, step=0.1, label="Temperature")
        max_new_tokens_slider = gr.Slider(1, 1000, 100, step=1, label="Max New Tokens")
        language_dropdown = gr.Dropdown(choices=["en", "ru"], value="en", label="Speech Language")

    send_btn.click(generate_response, [chatbot, txt, temperature_slider, max_new_tokens_slider, language_dropdown], [chatbot, audio_output])
    txt.submit(generate_response, [chatbot, txt, temperature_slider, max_new_tokens_slider, language_dropdown], [chatbot, audio_output])
    clear_btn.click(clear_chat, outputs=[chatbot, txt])

# Launch the interface
if __name__ == "__main__":
    demo.launch()