Alfred-40B-0723 was trained on AWS Sagemaker and is available on HuggingFace and soon on AWS Jumpstart for Foundation Models. By offering Alfred-40B-0723 as an open-source model, LightOn aims to foster collaboration and innovation within the Generative AI community, encouraging developers, researchers, and organizations to contribute to Alfred's further development.
Key takeaways:
- Alfred-40B-0723 is a new open-source Language Model (LLM) based on Falcon-40B, designed to integrate Generative AI into business workflows.
- Alfred-40B-0723 can perform tasks such as prompt engineering, no-code application development, and traditional LLM tasks.
- Alfred-40B-0723 is the first finetuned version of Falcon obtained with Reinforcement Learning from Human Feedback, and powers Paradigm, the Generative AI platform for the enterprise.
- Alfred-40B-0723 has been trained on AWS Sagemaker, is available on HuggingFace, and will be available on AWS Jumpstart for Foundation Models.