Access to Meta Model LLama 3.3 for GKE

Beginners
1

Hi All,

I cannot make my GKE cluster to use meta-llama/Meta-Llama-3.3-70B-Instruct even the web page said:
"
meta-llama/Llama-3.3-70B-Instruct"
You have been granted access to this model
"
with:
cat download_model.py
"
import os
from huggingface_hub import hf_hub_download

token = os.getenv(“HF_TOKEN”)
print(token)
hf_hub_download(repo_id=“meta-llama/Meta-Llama-3.3-70B-Instruct”, filename=“config.json”, token=token)"
"
The HF_TOKEN is a read token. and I change it many time… cannot be posted here.

What this can be?

Thanks a lot

1 Like
2

hf_hub_download is a function for downloading a single file. Let’s use snapshot_download. The rest is correct.

3

Have you verified that your HF_TOKEN has the necessary permissions to access the model? Also, ensure your GKE environment has internet access and correct Python dependencies installed. You could try testing the download script locally to isolate the issue.

1 Like