Loading a model is going to HuggingFace

[ShakunBaichoo]

Hi!
The code to load the GREAT model is going to HuggingFace to download the model. I have set the path as prescribed. I would like to ensure that it loads it locally as our cluster env does not have access to Internet.

Here is the code I am using:
import os
import sys
sys.path.append('tabularfm')

from tabularfm.utils.cli import get_config, create_model, create_model_config
from tabularfm.utils.processing import load_model_weights
from transformers import AutoTokenizer

configs

model_type = 'great'
config_path = 'configs/great.yaml'

data_path = # kaggle path

data_path = 'datasets/gittables_v4' # gittables
finetune_path = 'models/great_gittables'

load and create config

configs = get_config(config_path)
model_config_finetune = create_model_config(data_path, configs, model_type, config_type = "finetune")
model_config_finetune['pretrained_llm'] = os.path.join(finetune_path, 'weights')

load model

great_model = create_model(model_type, model_config_finetune)

Kindly advise.

Thanks,
Shkauntala