mirror of
https://github.com/k2-fsa/icefall.git
synced 2025-08-10 02:22:17 +00:00
24 lines
886 B
Markdown
24 lines
886 B
Markdown
# Introduction
|
|
|
|
This recipe includes ASR models (zipformer, zipformer-hat, zipformer-hat-lid) trained and evaluated on SEAME dataset.
|
|
The SEAME corpora is Singaporean Codeswitched English and Mandarin.
|
|
|
|
This corpus comes defined with a training split and two development splits:
|
|
|
|
train -- A mix of codeswitched, Mandarin and Singaporean English
|
|
dev_sge -- A set of primarily Singaporean English though there is codeswitching
|
|
dev_man -- A set of primarily Mandarin though there is also some codeswitching
|
|
|
|
|
|
[./RESULTS.md](./RESULTS.md) contains the latest results.
|
|
|
|
# Zipformer-hat
|
|
|
|
Zipformer with hybrid autoregressive transducer (HAT) loss https://arxiv.org/abs/2003.07705
|
|
see https://github.com/k2-fsa/icefall/pull/1291
|
|
|
|
# Zipformer-hat-lid
|
|
|
|
Zipformer-hat with auxiliary LID encoder and blank sharing for synchronization between ASR and LID as described here (will be shared soon)
|
|
|