You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
PaddleSpeech/examples/other/rhy
Hui Zhang 2376c14d7b
fix ckpt
1 year ago
..
conf Revised structure of rhythm prediction, test=tts 2 years ago
data Revised structure of rhythm prediction, test=tts 2 years ago
local refactor rhy 1 year ago
README.md Update README.md 2 years ago
path.sh Added rhythm prediction function, test=tts 2 years ago
run.sh fix ckpt 1 year ago

README.md

Prosody Prediction with CSMSC and AISHELL-3

Get Started

Data Preprocessing

./run.sh --stage 0 --stop-stage 0

Model Training

./run.sh --stage 1 --stop-stage 1

Testing

./run.sh --stage 2 --stop-stage 2

Prosody Prediction

./run.sh --stage 3 --stop-stage 3

Pretrained Model

The pretrained model can be downloaded here:

ernie-1.0_aishellcsmsc_ckpt_1.3.0.zip

And you should put it into exp/${YOUREXP}/checkpoints folder.

Rhythm mapping

Four punctuation marks are used to denote the rhythm marks respectively:

ryh_token csmsc aishll3
% #1 %
` #2
~ #3
$ #4 $

Prediction Results

#1 #2 #3 #4
Precision 0.90 0.66 0.91 0.90
Recall 0.92 0.62 0.83 0.85
F1 0.91 0.64 0.87 0.87