🦙 OpenLLaMA is an open-source project aiming to recreate Meta's LLaMA language models for commercial use. It’s now available for free on GitHub



Berkeley AI Research is using the RedPajama dataset to develop OpenLLaMA, and they have released an early version with 13-billion parameters.



The team believes that the fully trained OpenLLaMA will compete with Meta's original models in performance.



What does it mean?
Free and open-source model based on Meta's data, is available and competing with ChatGPT and Bard.



Github: https://github.com/openlm-research/open_llama



Hugging Face:
https://huggingface.co/openlm-research/open_llama_13b