Added Cerebras LLM

[sk5268]

Purpose

Cerebras with its 3rd generation Wafer-Scale-Engine (WSE-3) offers inference speed of ~2000 TPS.
Hence I added cerebras api to the list of available llms using langchain-cerebras.

Usage

1. Import the llm class:

    from openagi.llms.cerebras import CerebrasModel

2. Set the ENV variables:

    os.environ['CEREBRAS_API_KEY'] = "YOUR_CEREBRAS_API_KEY"
    os.environ['Cerebras_MODEL'] = "llama-3.3-70b"
    os.environ['Cerebras_TEMP'] = "0.5"

3. Configre the LLM

    config = CerebrasModel.load_from_env_config()
    llm = CerebrasModel(config=config)

Note

I tested this tool by replacing gemini( from openagi.llms.gemini import GeminiModel )
with cerebras (from openagi.llms.cerebras import CerebrasModel) in the example notebook example/blog_post.py.

Everything worked as expected, same as using GeminiModel.

---

References

1. Cerebras AI
2. Langchain Cerebras Integration

[sk5268]

Updated Doc-string format