Accessing SciBERT Model

Beginners
1

Hi everyone,
I’m a complete beginner in using the Huggingface model. I’m trying to use the SciBERT model in Google Colab for an ablation study this way:

model_name = "allenai/scibert_base_uncased/tree/main"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels=2)

I am using a secret key in my colab session with ‘read’ access. But getting this error:

HFValidationError: Repo id must be in the form ‘repo_name’ or ‘namespace/repo_name’: ‘allenai/scibert_base_uncased/tree/main’. Use repo_type argument if needed.

Can someone please specify what am I doing wrong here and how can I access the model? Also, any resource on how to do ablation on text data would be really appreciated. Thanks in advance.

2

I can not find the exact model but here is a possible fix.

model_name = "allenai/scibert_scivocab_uncased"

You dont need to pass /tree/main in the string for the model name.