Model support: wav2vec2

[harryfei]

wav2vec 2.0 is a framework for self-supervised learning of speech representations. It can be used as fundamental model for other audio tasks.

• Introduction: https://huggingface.co/docs/transformers/v4.46.2/en/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer
• github: https://github.com/facebookresearch/fairseq/blob/main/fairseq/models/wav2vec/wav2vec2.py
• huggingface transformer implementation: https://github.com/huggingface/transformers/tree/v4.46.2/src/transformers/models/wav2vec2
• huggingface model: https://huggingface.co/facebook/wav2vec2-base-960h
• paper: https://arxiv.org/pdf/2006.11477