ESPnet2 ASR model
espnet/brianyan918_aesrc2020_asr_conformer
This model was trained by Brian Yan using the aesrc2020 recipe in espnet.
Demo: How to use in ESPnet2
cd espnet
pip install -e .
cd egs2/aesrc2020/asr1
./run.sh --skip_data_prep false --skip_train true --download_model espnet/brianyan918_aesrc2020_asr_conformer
RESULTS
Environments
- date:
Sat Aug 20 06:55:57 EDT 2022
- python version:
3.8.13 (default, Mar 28 2022, 11:38:47) [GCC 7.5.0]
- espnet version:
espnet 202207
- pytorch version:
pytorch 1.8.1
- Git hash:
c892feb2ba248c85b683bf3cdef6c8f7ce85449a
- Commit date:
Thu Aug 18 11:54:56 2022 -0400
asr_train_asr_raw_en_bpe5000_sp
WER
dataset |
Snt |
Wrd |
Corr |
Sub |
Del |
Ins |
Err |
S.Err |
decode_asr_asr_model_valid.acc.ave/test |
15299 |
146813 |
94.8 |
4.7 |
0.6 |
0.7 |
5.9 |
31.9 |
CER
dataset |
Snt |
Wrd |
Corr |
Sub |
Del |
Ins |
Err |
S.Err |
decode_asr_asr_model_valid.acc.ave/test |
15299 |
708064 |
97.9 |
1.2 |
0.9 |
0.7 |
2.8 |
31.9 |
TER
dataset |
Snt |
Wrd |
Corr |
Sub |
Del |
Ins |
Err |
S.Err |
decode_asr_asr_model_valid.acc.ave/test |
15299 |
165136 |
93.2 |
4.3 |
2.6 |
0.6 |
7.4 |
31.9 |