The procedure of downloading Llama

[XiongweiWu]

Hi, can u tell me the procedure of downloading the original weights of Llama-13b-hf?

According to your instruction, I fill in the form in their website, however, they has not replied me for nearly one week. I find the original weight of Llama-13b from this website and git pull the weights.

However, when I tried to convert the Llama into hf as suggested in https://huggingface.co/docs/transformers/main/model_doc/llama , one error reports that No such file or directory: 'llama-13b/13B/params.json'

I am very new to this domain, and can u tell me the correct procedure to obtain the Llama-13b-hf?

[WeileiZeng]

use absolute path for the folder /path/to/llama-13b/13B

[XiongweiWu]

@WeileiZeng I know how to config the dataset, the question is how to download the original model and convert it into hugging face format, since I follow the official script but the error reports. However, I obtain the hf model from the third party repo.

[birchmi]

@WeileiZeng I know how to config the dataset, the question is how to download the original model and convert it into hugging face format, since I follow the official script but the error reports. However, I obtain the hf model from the third party repo.

Hi, I also encountered the same problem. Could you please paste the download link of hf model?

[WeileiZeng]

llamA weights

magnet:?xt=urn:btih:ZXXDAUWYLRUXXBHUYEMS6Q5CE5WA3LVA&dn=LLaMA

[XiongweiWu]

@WeileiZeng I know how to config the dataset, the question is how to download the original model and convert it into hugging face format, since I follow the official script but the error reports. However, I obtain the hf model from the third party repo.

Hi, I also encountered the same problem. Could you please paste the download link of hf model?

https://huggingface.co/decapoda-research/llama-13b-hf