Search code examples
pythonnlpgoogle-colaboratoryhuggingface-transformers

Downloading huggingface pre-trained models

Once I have downloaded a pre-trained model on a Colab Notebook, it disappears after I reset the notebook variables. Is there a way I can download the model to use it for a second occasion?

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
Solution

  • Mount your google drive:

    from google.colab import drive
drive.mount('/content/drive')

    Do your stuff and save your models:

    from transformers import BertTokenizer

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
tokenizer.save_pretrained('/content/drive/My Drive/tokenizer/')

    Reload it in a new session:

    tokenizer2 = BertTokenizer.from_pretrained('/content/drive/My Drive/tokenizer/')