I am trying to load the Asclepius-7B model using Transformers and the following code.
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("starmpcc/Asclepius-Llama2-7B")
model = AutoModel.from_pretrained("starmpcc/Asclepius-Llama2-7B")
The loading does not finish however and the process is killed with the error message / warning:
UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown
I am trying to load it on my MacBook Pro with 8GB memory without GPU. My guess is, it is a memory problem. Is there a way I can use it anyway on my local computer?
If not, can I use something like a virtual machine to get access to more memory?
Thanks in advance