Model Speed Comparison Cookbook

You can also check this cookbook in colab here

Overview

In this section, we’ll explore the speed of different AI models by creating instances of ChatAgent and measuring their response times to user prompts.

Installation

Ensure you have CAMEL AI installed in your Python environment:

!pip install "camel-ai==0.2.9"

Setting Up API Keys

You’ll need to set up your API keys for OpenAI and SambaNova.

import os
from getpass import getpass

# Prompt for the API key securely
openai_api_key = getpass('Enter your OpenAI API key: ')
os.environ["OPENAI_API_KEY"] = openai_api_key

sambanova_api_key = getpass('Enter your SambaNova API key: ')
os.environ["SAMBA_API_KEY"] = sambanova_api_key

Enter your OpenAI API key: ··········
Enter your SambaNova API key: ··········

import time
import matplotlib.pyplot as plt
from camel.agents import ChatAgent
from camel.configs import SambaCloudAPIConfig, ChatGPTConfig
from camel.messages import BaseMessage
from camel.models import ModelFactory
from camel.types import ModelPlatformType, ModelType

# Create model instances
def create_models():
    model_configs = [
        (ModelPlatformType.OPENAI, ModelType.GPT_4O_MINI, ChatGPTConfig(temperature=0.0, max_tokens=2000), "OpenAI GPT-4O Mini"),
        (ModelPlatformType.OPENAI, ModelType.GPT_4O, ChatGPTConfig(temperature=0.0, max_tokens=2000), "OpenAI GPT-4O"),
        # NOTE: OpenAI O1 model requires additional additional credentials
        (ModelPlatformType.OPENAI, ModelType.O1_PREVIEW, ChatGPTConfig(temperature=0.0), "OpenAI O1 Preview"),
        (ModelPlatformType.SAMBA, "Meta-Llama-3.1-8B-Instruct", SambaCloudAPIConfig(temperature=0.0, max_tokens=2000), "SambaNova Llama 8B"),
        (ModelPlatformType.SAMBA, "Meta-Llama-3.1-70B-Instruct", SambaCloudAPIConfig(temperature=0.0, max_tokens=2000), "SambaNova Llama 70B"),
        (ModelPlatformType.SAMBA, "Meta-Llama-3.1-405B-Instruct", SambaCloudAPIConfig(temperature=0.0, max_tokens=2000), "SambaNova Llama 405B")
    ]

    models = [(ModelFactory.create(model_platform=platform, model_type=model_type, model_config_dict=config.as_dict(), url="https://api.sambanova.ai/v1" if platform == ModelPlatformType.SAMBA else None), name)
              for platform, model_type, config, name in model_configs]
    return models

# Define messages
def create_messages():
    sys_msg = BaseMessage.make_assistant_message(role_name="Assistant", content="You are a helpful assistant.")
    user_msg = BaseMessage.make_user_message(role_name="User", content="Tell me a long story.")
    return sys_msg, user_msg

# Initialize ChatAgent instances
def initialize_agents(models, sys_msg):
    return [(ChatAgent(system_message=sys_msg, model=model), name) for model, name in models]

# Measure response time for a given agent
def measure_response_time(agent, message):
    start_time = time.time()
    response = agent.step(message)
    end_time = time.time()
    tokens_per_second = response.info['usage']["completion_tokens"] / (end_time - start_time)
    return tokens_per_second

# Visualize results
def plot_results(model_names, tokens_per_sec):
    plt.figure(figsize=(10, 6))
    plt.barh(model_names, tokens_per_sec, color='skyblue')
    plt.xlabel("Tokens per Second")
    plt.title("Model Speed Comparison: Tokens per Second")
    plt.gca().invert_yaxis()
    plt.show()

# Main execution
models = create_models()
sys_msg, user_msg = create_messages()
agents = initialize_agents(models, sys_msg)

# Measure response times and collect data
model_names = []
tokens_per_sec = []

for agent, model_name in agents:
    model_names.append(model_name)
    tokens_per_sec.append(measure_response_time(agent, user_msg))

# Visualize the results
plot_results(model_names, tokens_per_sec)

/usr/local/lib/python3.10/dist-packages/camel/models/openai_model.py:107: UserWarning: Warning: You are using an O1 model (O1_MINI or O1_PREVIEW), which has certain limitations, reference: `https://platform.openai.com/docs/guides/reasoning`.
  warnings.warn(