Spaces:
Sleeping
Sleeping
File size: 2,731 Bytes
43b5bef c1e5d4c 518be16 c1e5d4c 43b5bef c1e5d4c cf38aa5 43b5bef c1e5d4c 43b5bef c1e5d4c 518be16 c1e5d4c 518be16 77ac272 c1e5d4c 6617dfe c1e5d4c 6617dfe 518be16 77ac272 c1e5d4c 6617dfe c1e5d4c c29137d c1e5d4c 518be16 c1e5d4c 518be16 c1e5d4c 43b5bef c1e5d4c 43b5bef c1e5d4c 43b5bef c1e5d4c |
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 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 |
import gradio as gr
import os
import requests
import threading
from typing import List, Dict, Any
# Get the Hugging Face API key from Spaces secrets
HF_API_KEY = os.getenv("HF_API_KEY")
# Model endpoints configuration
MODEL_ENDPOINTS = {
"Qwen2.5-72B-Instruct": "https://api-inference.huggingface.co/models/Qwen/Qwen2.5-72B-Instruct",
"Llama3.3-70B-Instruct": "https://api-inference.huggingface.co/models/meta-llama/Llama-3.3-70B-Instruct",
"Qwen2.5-Coder-32B-Instruct": "https://api-inference.huggingface.co/models/Qwen/Qwen2.5-Coder-32B-Instruct",
}
def query_model(model_name: str, messages: List[Dict[str, str]]) -> str:
"""Query a single model with the chat history"""
endpoint = MODEL_ENDPOINTS[model_name]
headers = {
"Authorization": f"Bearer {HF_API_KEY}",
"Content-Type": "application/json"
}
# Format the prompt according to each model's requirements
prompt = "\n".join([f"{msg['role']}: {msg['content']}" for msg in messages])
payload = {
"inputs": prompt,
"parameters": {
"max_tokens": 1024,
"temperature": 0.7,
"stop_sequences": ["\nUser:", "\nAssistant:", "###"]
}
}
try:
response = requests.post(endpoint, json=payload, headers=headers)
response.raise_for_status()
return response.json()[0]['generated_text']
except Exception as e:
return f"Error from {model_name}: {str(e)}"
def respond(message: str, history: List[List[str]]) -> str:
"""Handle chat responses from all models"""
# Prepare messages in OpenAI format
messages = [{"role": "user", "content": message}]
# Create threads for concurrent model queries
threads = []
results = {}
def get_model_response(model_name):
results[model_name] = query_model(model_name, messages)
for model_name in MODEL_ENDPOINTS:
thread = threading.Thread(target=get_model_response, args=(model_name,))
thread.start()
threads.append(thread)
# Wait for all threads to complete
for thread in threads:
thread.join()
# Format responses from all models
responses = []
for model_name, response in results.items():
responses.append(f"**{model_name}**:\n{response}")
return "\n\n".join(responses)
# Create the Gradio interface
chat_interface = gr.ChatInterface(
respond,
title="Multi-LLM Collaboration Chat",
description="A group chat with Qwen2.5-72B, Llama3.3-70B, and Qwen2.5-Coder-32B",
examples=["How can I optimize Python code?", "Explain quantum computing basics"],
theme="soft"
)
if __name__ == "__main__":
chat_interface.launch(share=True)
|