Swarm Models

Swarm Models provides a unified, secure, and highly scalable interface for interacting with multiple LLM and multi-modal APIs across different providers. It is built to streamline your API integrations, ensuring production-grade reliability and robust performance.

Key Features:

• Multi-Provider Support: Integrate seamlessly with APIs from OpenAI, Anthropic, Azure, and more.

• Enterprise-Grade Security: Built-in security protocols to protect your API keys and sensitive data, ensuring compliance with industry standards.

• Lightning-Fast Performance: Optimized for low-latency and high-throughput, Swarm Models delivers blazing-fast API responses, suitable for real-time applications.

• Ease of Use: Simplified API interaction with intuitive .run(task) and __call__ methods, making integration effortless.

• Scalability for All Use Cases: Whether it's a small script or a massive enterprise-scale application, Swarm Models scales effortlessly.

• Production-Grade Reliability: Tested and proven in enterprise environments, ensuring consistent uptime and failover capabilities.

---

Onboarding

Swarm Models simplifies the way you interact with different APIs by providing a unified interface for all models.

1. Install Swarm Models

$ pip3 install -U swarm-models

2. Set Your Keys

OPENAI_API_KEY="your_openai_api_key"
GROQ_API_KEY="your_groq_api_key"
ANTHROPIC_API_KEY="your_anthropic_api_key"
AZURE_OPENAI_API_KEY="your_azure_openai_api_key"

3. Initialize a Model

Import the desired model from the package and initialize it with your API key or necessary configuration.

from swarm_models import YourDesiredModel

model = YourDesiredModel(api_key='your_api_key', *args, **kwargs)

4. Run Your Task

Use the .run(task) method or simply call the model like model(task) with your task.

task = "Define your task here"
result = model.run(task)

# Or equivalently
#result = model(task)

5. Enjoy the Results

print(result)

---

Full Code Example

from swarm_models import OpenAIChat
import os

# Get the OpenAI API key from the environment variable
api_key = os.getenv("OPENAI_API_KEY")

# Create an instance of the OpenAIChat class
model = OpenAIChat(openai_api_key=api_key, model_name="gpt-4o-mini")

# Query the model with a question
out = model(
   "What is the best state to register a business in the US for the least amount of taxes?"
)

# Print the model's response
print(out)

---

TogetherLLM Documentation

The TogetherLLM class is designed to simplify the interaction with Together's LLM models. It provides a straightforward way to run tasks on these models, including support for concurrent and batch processing.

Initialization

To use TogetherLLM, you need to initialize it with your API key, the name of the model you want to use, and optionally, a system prompt. The system prompt is used to provide context to the model for the tasks you will run.

Here's an example of how to initialize TogetherLLM:

import os
from swarm_models import TogetherLLM

model_runner = TogetherLLM(
    api_key=os.environ.get("TOGETHER_API_KEY"),
    model_name="meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
    system_prompt="You're Larry fink",
)

Running Tasks

Once initialized, you can run tasks on the model using the run method. This method takes a task string as an argument and returns the response from the model.

Here's an example of running a single task:

task = "How do we allocate capital efficiently in your opinion Larry?"
response = model_runner.run(task)
print(response)

Running Multiple Tasks Concurrently

TogetherLLM also supports running multiple tasks concurrently using the run_concurrently method. This method takes a list of task strings and returns a list of responses from the model.

Here's an example of running multiple tasks concurrently:

tasks = [
    "What are the top-performing mutual funds in the last quarter?",
    "How do I evaluate the risk of a mutual fund?",
    "What are the fees associated with investing in a mutual fund?",
    "Can you recommend a mutual fund for a beginner investor?",
    "How do I diversify my portfolio with mutual funds?",
]
responses = model_runner.run_concurrently(tasks)
for response in responses:
    print(response)

Enterprise-Grade Features

1. Security: API keys and user data are handled with utmost care, utilizing encryption and best security practices to protect your sensitive information.

2. Production Reliability: Swarm Models has undergone rigorous testing to ensure that it can handle high traffic and remains resilient in enterprise-grade environments.

3. Fail-Safe Mechanisms: Built-in failover handling to ensure uninterrupted service even under heavy load or network issues.

4. Unified API: No more dealing with multiple SDKs or libraries. Swarm Models standardizes your interactions across providers like OpenAI, Anthropic, Azure, and more, so you can focus on what matters.

---

Available Models

Model Name	Import Path
BaseLLM	from swarm_models.base_llm import BaseLLM
BaseMultiModalModel	from swarm_models.base_multimodal_model import BaseMultiModalModel
GPT4VisionAPI	from swarm_models.gpt4_vision_api import GPT4VisionAPI
HuggingfaceLLM	from swarm_models.huggingface import HuggingfaceLLM
LayoutLMDocumentQA	from swarm_models.layoutlm_document_qa import LayoutLMDocumentQA
llama3Hosted	from swarm_models.llama3_hosted import llama3Hosted
LavaMultiModal	from swarm_models.llava import LavaMultiModal
Nougat	from swarm_models.nougat import Nougat
OpenAIEmbeddings	from swarm_models.openai_embeddings import OpenAIEmbeddings
OpenAITTS	from swarm_models.openai_tts import OpenAITTS
GooglePalm	from swarm_models.palm import GooglePalm as Palm
Anthropic	from swarm_models.popular_llms import Anthropic as Anthropic
AzureOpenAI	from swarm_models.popular_llms import AzureOpenAILLM as AzureOpenAI
Cohere	from swarm_models.popular_llms import CohereChat as Cohere
OctoAIChat	from swarm_models.popular_llms import OctoAIChat
OpenAIChat	from swarm_models.popular_llms import OpenAIChatLLM as OpenAIChat
OpenAILLM	from swarm_models.popular_llms import OpenAILLM as OpenAI
Replicate	from swarm_models.popular_llms import ReplicateChat as Replicate
QwenVLMultiModal	from swarm_models.qwen import QwenVLMultiModal
FireWorksAI	from swarm_models.popular_llms import FireWorksAI
Vilt	from swarm_models.vilt import Vilt
TogetherLLM	from swarm_models.together_llm import TogetherLLM
LiteLLM	from swarm_models.lite_llm_model import LiteLLM
OpenAIFunctionCaller	from swarm_models.openai_function_caller import OpenAIFunctionCaller
OllamaModel	from swarm_models.ollama_model import OllamaModel
GroundedSAMTwo	from swarm_models.sam_two import GroundedSAMTwo

---

Support & Contributions

• Documentation: Comprehensive guides, API references, and best practices are available in our official Documentation.
• GitHub: Explore the code, report issues, and contribute to the project via our GitHub repository.

---

License

Swarm Models is released under the MIT License.

---

Todo

• Add cohere models command r
• Add gemini and google ai studio
• Integrate ollama extensively