Skip to content

Latest commit

 

History

History
28 lines (19 loc) · 4.34 KB

pretrained_models.md

File metadata and controls

28 lines (19 loc) · 4.34 KB
Raw

Pretrained Models in WeNet

Model Types

We provide two types of pretrained model in WeNet to facilitate users with different requirements.

  1. Checkpoint Model, with suffix .pt, the model trained and saved as checkpoint by WeNet python code, you can reproduce our published result with it, or you can use it as checkpoint to continue.

  2. Runtime Model, with suffix .zip, you can directly use runtime model in our x86 or android runtime, the runtime model is export by Pytorch JIT on the checkpoint model. Two kinds of runtime models are provided:

    • x86, server model, typically big.
    • android, on-device model, typically small and been quantized.

Model License

The pretrained model in WeNet follows the license of it's corresponding dataset. For example, the pretrained model on LibriSpeech follows CC BY 4.0, since it is used as license of the LibriSpeech dataset, see http://openslr.org/12/.

Model List

Here is a list of the pretrained models on different datasets. The model structure, model size, and download link are given.

Datasets Languages Checkpoint Model Runtime Model(x86) Runtime Model(android)
aishell CN Conformer/174M/Download U2 Transformer/127M/Download U2 Transformer/38M/Download
aishell2 CN U2++ Conformer/187M/Download U2 Transformer/130M/Download U2 Transformer/39M/Download
gigaspeech EN U2++ Conformer/472M/Download U2++ Conformer/507M/Download U2++ Transformer/51M/Download
librispeech EN Conformer/481M/Download U2++ Conformer/199M/Download
multi_cn CN U2 Conformer/193M/Download U2 Conformer/130M/Download U2 Conformer/65M/Download