No space left on device when deploying

[fredrikst]

I'm using HuggingfaceModel to create the model and deploy an endpoint. The instance type is ml.g5.x2large, which clearly states 450GB storage space in the specs. The model size is around 30GB.

What am I missing here, do I need to specify a certain cache_dir?

[pottieral]

Hi, I am facing the same problem, were you able to solve it?

[Wingjam]

Im also facing the same issue with the following code:

from sagemaker.huggingface import HuggingFaceModel

huggingface_model = HuggingFaceModel(
    model_data=s3_model_uri,
    role=role,
    transformers_version="4.26.0",
    pytorch_version="1.13.1",
    py_version="py39",
    entry_point="inference.py",
)

predictor = huggingface_model.deploy(
    initial_instance_count=1,
    instance_type="ml.g4dn.8xlarge",
)

print(f"Endpoint Name: {predictor.endpoint_name}")

Got this error: OSError: [Errno 28] No space left on device