I’m attempting to do inference using the latest “robust” model, roughly following the steps laid out in inspired by this notebook
When attempting to load in the Tokenizer for wav2vec2-large-robust, I have been encountering the following issue:
OSError: Can't load tokenizer for 'facebook/wav2vec2-large-robust'. Make sure that:
- 'facebook/wav2vec2-large-robust' is a correct model identifier listed on 'https://huggingface.co/models'
- or 'facebook/wav2vec2-large-robust' is the correct path to a directory containing relevant tokenizer files
I tried AutoTokenizer, Wav2Vec2CTCTokenizer and Wav2Vec2Tokenizer, but none of those work.
When I use the same code, but swap out the name of the pretrained model to, say, “facebook/wav2vec2-base-960h” it works.
I tried retyping the names a few times, and recopied it directly from the website. I also recreated the issue in a Colab notebook, below.
Colab notebook replicating the issue: https://colab.research.google.com/drive/1tZ3oEtkg9_74nCLnnRq9lpLktyfAAS6I?usp=sharing
Anyone got any suggestions for what I should try?