alvanlii/wav2vec2-BERT-cantonese

Automatic Speech Recognition·alvanlii· 532.7K· 6

transformers apache-2.0 608.4M params dataset:mozilla-foundation/common_voice_16_0arxiv:2201.02419license:apache-2.0region:us

This model is a fine-tuned version of facebook/w2v-bert-2.0. This has a CER of 10.27 on Common Voice 16 (yue) test set (without punctuations).

# pull & run locally
pip install mlforge-sdk && mlforge pull alvanlii/wav2vec2-BERT-cantonese

Model details

Task

Automatic Speech Recognition

Provider

alvanlii

Framework

transformers

Parameters

608.4M

Size

27 GB

License

apache-2.0

Downloads

532.7K

Likes

Paper

arXiv:2201.02419

Updated

2024-04-05

This model is a fine-tuned version of facebook/w2v-bert-2.0. This has a CER of 10.27 on Common Voice 16 (yue) test set (without punctuations).