Model reference · open weights
wavlm-large is an open-weight embedding model from microsoft. wavlm-large (BF16) weighs 1.3 GB; the smallest configuration that runs it is RTX 3060 12 GB.
What it is
| Released by | microsoft |
|---|---|
| Released | 2022-03-02 |
| VRAM | 1.3 GB for the weights |
What it runs on
| Card | Runs | Memory |
|---|---|---|
| RTX 3060 12 GB | fits | 11.6 GB |
| RTX 4060 Ti 16 GB | fits | 15.4 GB |
| RTX 3090 24 GB | fits | 23.4 GB |
| RTX 4090 24 GB | fits | 23.4 GB |
| RTX 5090 32 GB | fits | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
The large model pretrained on 16kHz sampled speech audio. When using the model, make sure that your speech input is also sampled at 16kHz.
Note: This model does not have a tokenizer as it was pretrained on audio alone. In order to use this model speech recognition, a tokenizer should be created and the model should be fine-tuned on labeled text data. Check out this blog for more in-detail explanation of how to fine-tune the model.
The model was pre-trained on:
Paper: WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing
Authors: Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei
Abstract Self-supervised learning (SSL) achieves great success in speech recognition, while limited exploration has been attempted for other speech processing tasks. As speech signal contains multi-faceted information including speaker identity, paralinguistics, spoken content, etc., learning universal representations for all speech tasks is challenging. In this paper, we propose a new pre-trained model, WavLM, to solve full-stack downstream speech tasks. WavLM is built based on the HuBERT framework, with an emphasis on both spoken content modeling and speaker identity preservation. We first equip the Transformer structure with gated relative position bias to improve its capability on recognition tasks. For better speaker discrimination, we propose an utterance mixing training strategy, where additional overlapped utterances are created unsupervisely and incorporated during model training. Lastly, we scale up the training dataset from 60k hours to 94k hours. WavLM Large achieves state-of-the-art performance on the SUPERB benchmark, and brings significant improvements for various speech processing tasks on their representative benchmarks.
The original model can be found under https://github.com/microsoft/unilm/tree/master/wavlm.
This is an English pre-trained speech model that has to be fine-tuned on a downstream task like speech recognition or audio classification before it can be used in inference. The model was pre-trained in English and should therefore perform well only in English. The model has been shown to work well on the SUPERB benchmark.
Note: The model was pre-trained on phonemes rather than characters. This means that one should make sure that the input text is converted to a sequence of phonemes before fine-tuning.
To fine-tune the model for speech recognition, see the official speech recognition example.
To fine-tune the model for speech classification, see the official audio classification example.
TODO
TODO
The model was contributed by cywang and patrickvonplaten.
The official license can be found here
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.