I’m running miniconda on GCP debian DeepLearning VM.
I’m basically trying to run this:
embed_model = HuggingFaceEmbeddings(
model_name=embed_model_id,
model_kwargs={'device': device},
encode_kwargs={'device': device, 'batch_size': 32}
)
and have this error:
permission denied [Errno 13] Permission denied: '/root/google_vm_config.lock'
google my ass out. check other stackoverflow questions.
2
Answers
type
sudo -s
to get into root user Thensudo chown -R user:user /root/google_vm_config.lock
where user is you.
In one of the GitHub discussions I found a working solution. You need to add
GOOGLE_VM_CONFIG_LOCK_FILE
into list of ignore env vars in the functionto_be_ignored
located in this filelib/python3.10/site-packages/bitsandbytes/cuda_setup/env_vars.py
Link for reference.