Meta has announced the launch of Llama 2.0, a new version of its open source generative language model. Llama 2.0 is touted as more powerful than its previous version and ranks among the top-performing models. It offers a range of models with different parameters (7B, 13B, and 70B) as well as pre-trained variations dedicated to chat completion.
In essence, it can generate longer and more complex text, and it is capable of following more intricate instructions. This makes it a more versatile tool that can be used for a broader range of tasks.
Llama 2.0 is available for free on Hugging Face or on the project’s GitHub website accessible here:
GitHub – facebookresearch/llama: Inference code for LLaMA models
However, an upstream request must be made at the following link to obtain the necessary URL for downloading the different models: