Back to all updates

over 1 year ago

GPT4o: Code & Conquer Hackathon - Resources & Security Reminder

Hello Developers,

The GPT4o: Code & Conquer Hackathon is now officially live! 

To get you started, here are your resources:

1. LLaMA Model (Meta-Llama/Llama-3.2-3B) hosted on Hugging Face: LLaMA Model Inference Link

 Here’s a code snippet to quickly call the LLaMA model:

import requests

# API endpoint for the LLaMA model hosted on Hugging Face
API_URL = "https://l7sol6qs4x9pu9j7.us-east-1.aws.endpoints.huggingface.cloud"
headers = {
    "Accept": "application/json",
    "Content-Type": "application/json"
}

def query(payload):
    """
    Sends a POST request to the LLaMA model API with the given payload.

    Args:
        payload (dict): The payload containing inputs and parameters for the model.

    Returns:
        dict: The JSON response from the API.
    """
    response = requests.post(API_URL, headers=headers, json=payload)
    return response.json()

# Sample payload to make a query
output = query({
    "inputs": "Can you please let us know more details about your ",
    "parameters": {
        "top_p": 0.01,
        "temperature": 0.7,  # Adjust this value as needed for more creative or conservative responses
        "max_new_tokens": 150,
        "return_text": True,
        "return_full_text": True,
        "return_tensors": True,
        "clean_up_tokenization_spaces": True,
        "prefix": "#",
        "handle_long_generation": "hole"
    }
})

# Printing the output
print(output)

Feel free to use these resources or any other llm models of your choice to power your project.

Quick Reminders:

• Submission: Be sure to submit your projects on Devpost before the deadline.
• Guidance Available: Industry experts are available for mentorship via our Discord server: https://discord.gg/ETMYVf2BJd

Wishing you all the best in your hacking journey!

If you have any questions, I’m here to help.

Best regards,
Yash