You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
At the deserialization of a model, the tokenizer gets loaded from huggingface (we're using TransformerWordEmbeddings). This throws an error when for instance using SCIBERT, because it is not supported in huggingface.
To Reproduce
Train a sequence tagger model using scibert. Then according to the docs: tagger = SequenceTagger("model_using_scibert"), which throws an errors since it is looking the scibert tokenizer in the huggingface lib.
Expected behavior
The tokenizer should not get loaded from huggingface if it is not available.
Environment (please complete the following information):
macOS Catalina
Version flair 0.5
The text was updated successfully, but these errors were encountered:
Describe the bug
At the deserialization of a model, the tokenizer gets loaded from huggingface (we're using TransformerWordEmbeddings). This throws an error when for instance using SCIBERT, because it is not supported in huggingface.
To Reproduce
Train a sequence tagger model using scibert. Then according to the docs: tagger = SequenceTagger("model_using_scibert"), which throws an errors since it is looking the scibert tokenizer in the huggingface lib.
Expected behavior
The tokenizer should not get loaded from huggingface if it is not available.
Environment (please complete the following information):
The text was updated successfully, but these errors were encountered: