Spaces:
Sleeping
Sleeping
File size: 2,003 Bytes
2f3b32c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
import os
import re
import gradio as gr
import edge_tts
import asyncio
import time
import tempfile
from huggingface_hub import InferenceClient
DESCRIPTION = """ # <center><b>JARVIS⚡</b></center>
### <center>A personal Assistant of Tony Stark for YOU
### <center>Currently It supports text input, But If this space completes 1k hearts than I starts working on Audio Input.</center>
"""
client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
system_instructions = "[INST] Answer as Real Jarvis JARVIS, Made by 'Tony Stark', Keep conversation very short, clear, friendly and concise."
async def generate(prompt):
generate_kwargs = dict(
temperature=0.6,
max_new_tokens=100,
top_p=0.95,
repetition_penalty=1,
do_sample=True,
seed=42,
)
formatted_prompt = system_instructions + prompt + "[/INST]"
stream = client.text_generation(
formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
output = ""
for response in stream:
output += response.token.text
communicate = edge_tts.Communicate(output)
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
tmp_path = tmp_file.name
await communicate.save(tmp_path)
yield tmp_path
with gr.Blocks(css="style.css") as demo:
gr.Markdown(DESCRIPTION)
with gr.Row():
user_input = gr.Textbox(label="Prompt")
input_text = gr.Textbox(label="Input Text", elem_id="important")
output_audio = gr.Audio(label="Audio", type="filepath",
interactive=False,
autoplay=True,
elem_classes="audio")
with gr.Row():
translate_btn = gr.Button("Response")
translate_btn.click(fn=generate, inputs=user_input,
outputs=output_audio, api_name="translate")
if __name__ == "__main__":
demo.queue(max_size=20).launch()
|