locally installed Llama2 7b hf

[georgejoy88]

Hi - i tried to load a locally installed llama2 7b hf rather than from the huggingface website but the gpu killed the operation. Please could I check if this is even possible?

[Infernaught]

What is the command you are running to load in the local model?

[georgejoy88]

i just replaced the huggingface destinationpath for the model with a local destinationpath?

zero_shot_config = yaml.safe_load(

"""

model_type: llm

base_model: meta-llama/Llama-2-7b-hf

[Infernaught]

Ahh, I see. Is this local model a pretrained adapter model or is it just the base llama-2-7b-hf that has been downloaded locally?