Error in running tokenize_and_split_data

Hey,

There seems to be a problem with the config module. Any suggestions?


AttributeError Traceback (most recent call last)
Cell In[28], line 5
3 tokenizer = AutoTokenizer.from_pretrained(model_name)
4 tokenizer.pad_token = tokenizer.eos_token
----> 5 train_dataset, test_dataset = tokenize_and_split_data(training_config, tokenizer)

File ~/utilities.py:111, in tokenize_and_split_data(training_config, tokenizer)
110 def tokenize_and_split_data(training_config, tokenizer):
→ 111 initialized_config = initialize_config_and_logging(training_config)
112 dataset_path = initialized_config[“datasets”][“path”]
113 use_hf = initialized_config[“datasets”][“use_hf”]

File ~/utilities.py:27, in initialize_config_and_logging(existing_config)
25 def initialize_config_and_logging(existing_config=None):
26 global global_config
—> 27 global_config = build_config(existing_config)
28 setup_logging(global_config)
29 logger.debug("Config: " + str(yaml.dump(global_config.as_dict())))

File ~/utilities.py:40, in build_config(existing_config)
37 def build_config(existing_config=None):
38 configs = [
39 # Using config library
—> 40 config.config_from_env(prefix=“LLAMA”, separator=“_”, lowercase_keys=True),
41 ]
43 if existing_config:
44 if isinstance(existing_config, dict):

AttributeError: module ‘config’ has no attribute ‘config_from_env’

Facing the same issue. Did someone find a solution to this?