About the LLaMA Format

[Albertchamberlain]

I see this sentence in your README file,does that mean i need convert the origin LLaMa?

Origin LLama dictionary structure like this 👇：

when i load the LLaMa Pretrain,there are something stucked，so i think i need convert the LLaMA model

[Albertchamberlain]

Does you use the LLaMA like this ？