Instructions to use facebook/esmfold_v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use facebook/esmfold_v1 with Transformers:
# Load model directly from transformers import AutoTokenizer, EsmForProteinFolding tokenizer = AutoTokenizer.from_pretrained("facebook/esmfold_v1") model = EsmForProteinFolding.from_pretrained("facebook/esmfold_v1") - Notebooks
- Google Colab
- Kaggle
Why is the special token for padding `A` in `special_tokens_map.json`?
#3
by ardhe-qb - opened
Hi,
Why is the special token for padding A in special_tokens_map.json? It seems like the padding token is <pad> in the vocabulary, in the default instantiation of EsmTokenizer, etc.
Is there a reason for this? It seems strange to add random Alanine's to pad a protein so that the whole batch is of the same size.
Thanks in advance for your reply!