meta-llama/Llama-2-70b-hf filling up my disk

Hello, I’m using
model = “meta-llama/Llama-2-70b-chat-hf”
tokenizer = AutoTokenizer.from_pretrained(model)
and it automatically downloads the model.
The model is going to filling up my disk and when the disk space is 0 the download ends. How should I solve it? I’m using a dedicated compute in Azure Machine Learning studio. It happens also in colab.