Has anyone tried to load the LLaMA 2

This might help

https://discord.com/channels/1059199217496772688/1131989723120926740/1133201476727083189

Thank you!

Instead of manually downloading the checkpoints, can I directly pull them with HuggingFaceLLM like model = ‘meta-llama/Llama-2-13b-chat-hf’?

Like here:
https://thealgorithmicminds.com/how-to-use-huggingface-to-use-llama-2-on-your-custom-machine-35713a2964de

As long as you have access from huggingface, yea I think so

You can create the model and tokenzier like that example, and then pass them in as kwargs to the huggingface llm. The example in the link above will download them automatically