# Inference Llama-2-13b not working
**URL:**
**Author:**  [@Tanuj](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FTanuj)%5C
**Post date:** [July 19, 2023, 4:44pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Finference-llama-2-13b-not-working%2F47375%2F1 "2023-07-19T16:44:10Z")
Following the [blog](https://huggingface.co/blog/llama2#inference), I was able to run inference on my gpu server for _meta-llama/Llama-2-7b-chat-hf_ as has been done in the tutorial. (Have been granted access to Llama-2 already)
Now I wish to run inference on _meta-llama/Llama-2-13b_. I keep getting the error:
**“OSError: meta-llama/Llama-2-13b does not appear to have a file named config.json. Checkout ‘[https://huggingface.co/meta-llama/Llama-2-13b/main](https://huggingface.co/meta-llama/Llama-2-13b/main)’ for available files.”**
Checking the mentioned repo for _config.json_, I see there actually isn’t one for _meta-llama/Llama-2-13b_ (and most other Llama2 models except _meta-llama/Llama-2-7b-chat-hf_).
Could the missing file(s) it be added please? Alternatively, if there’s another way to do a run the model privately, it would be great. Thanks!
---
**Author:**  [@0xmaddie](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2F0xmaddie)%5C
**Post date:** [July 19, 2023, 7:16pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Finference-llama-2-13b-not-working%2F47375%2F2 "2023-07-19T19:16:37Z")
I was able to load llama-2-13b-chat-hf with this Google Colab notebook, but inference failed for me as well due to a RuntimeError. I’m not sure if it’s entirely related to your problem, but maybe you’d like to take a look?
> **[Google Colaboratory](https://colab.research.google.com/github/0xmaddie/0xmaddie/blob/main/llama_debug.ipynb)**
---
**Author:**  [@Tanuj](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FTanuj)%5C
**Post date:** [July 20, 2023, 4:13pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Finference-llama-2-13b-not-working%2F47375%2F3 "2023-07-20T16:13:18Z")
Got the model working. Just need to use the models with ‘hf’ in the name.
---
**Author:**  [@krishnagarg09](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fkrishnagarg09)%5C
**Post date:** [July 21, 2023, 9:36pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Finference-llama-2-13b-not-working%2F47375%2F4 "2023-07-21T21:36:02Z")
Just to make it more clear, this runs perfectly fine:
**meta-llama/Llama-2-13b-hf**
---
**Author:**  [@flaviobrio](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fflaviobrio)%5C
**Post date:** [July 30, 2023, 8:47pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Finference-llama-2-13b-not-working%2F47375%2F5 "2023-07-30T20:47:05Z")
But it says that PRO license is required to use hf ones