This PR introduces the integration of multiple LLMs to enhance flexibility and adaptability:
Anthropic: Fully integrated and works seamlessly for generation. However, it currently lacks an embedding algorithm. To address this, OpenAI embeddings are being used as a workaround.
GPT4All: Integrated and running successfully locally. While it's slower compared to cloud-based models, it offers great flexibility for devices with varying computational power.
Future Improvements:
Plan to add support for Hugging Face embeddings to improve versatility and reduce reliance on specific APIs.
Key Features:
Multi-model support for generation and embeddings.
Adaptable to different computational capacities.
Clear structure for integrating new model options in the future.
Let me know your feedback, suggestions, or any potential improvements! 🚀
This PR introduces the integration of multiple LLMs to enhance flexibility and adaptability:
Future Improvements:
Key Features:
Let me know your feedback, suggestions, or any potential improvements! 🚀