Hubert Large Ll60k

🚀 Hubert-Large

Hubert-Large is a large-scale model pre-trained on 16kHz sampled speech audio. It offers a powerful solution for speech-related tasks. When using this model, ensure that your speech input is also sampled at 16kHz.

✨ Features

• Pretrained on 16kHz sampled speech audio from the Libri-Light dataset.
• Based on Facebook's Hubert approach.
• Can be fine-tuned for speech recognition tasks.

📦 Installation

No specific installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

Basic Usage

To fine-tune the model, refer to this blog for detailed information. Note that the class Wav2Vec2ForCTC has to be replaced by HubertForCTC.

Advanced Usage

For more advanced scenarios, you can explore different fine-tuning strategies and hyperparameters based on your specific requirements.

📚 Documentation

Model Information

Paper

• Paper
• Authors: Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed

Abstract

Self-supervised approaches for speech representation learning are challenged by three unique problems: (1) there are multiple sound units in each input utterance, (2) there is no lexicon of input sound units during the pre-training phase, and (3) sound units have variable lengths with no explicit segmentation. To deal with these three problems, we propose the Hidden-Unit BERT (HuBERT) approach for self-supervised speech representation learning, which utilizes an offline clustering step to provide aligned target labels for a BERT-like prediction loss. A key ingredient of our approach is applying the prediction loss over the masked regions only, which forces the model to learn a combined acoustic and language model over the continuous inputs. HuBERT relies primarily on the consistency of the unsupervised clustering step rather than the intrinsic quality of the assigned cluster labels. Starting with a simple k-means teacher of 100 clusters, and using two iterations of clustering, the HuBERT model either matches or improves upon the state-of-the-art wav2vec 2.0 performance on the Librispeech (960h) and Libri-light (60,000h) benchmarks with 10min, 1h, 10h, 100h, and 960h fine-tuning subsets. Using a 1B parameter model, HuBERT shows up to 19% and 13% relative WER reduction on the more challenging dev-other and test-other evaluation subsets.

Original Model

The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/hubert.

🔧 Technical Details

The technical details are mainly covered in the abstract and the paper, which explain the approach to deal with the unique problems in self-supervised speech representation learning and the performance on benchmarks.

📄 License

This model is released under the Apache 2.0 license.

⚠️ Important Note

This model does not have a tokenizer as it was pretrained on audio alone. In order to use this model for speech recognition, a tokenizer should be created and the model should be fine-tuned on labeled text data.

💡 Usage Tip

Check out this blog for more in-detail explanation of how to fine-tune the model.