[Hackathon 7th] 修复 g2p run.sh 中错误的命令 by megemini · Pull Request #3886 · PaddlePaddle/PaddleSpeech (original) (raw)


> bash run.sh 
Start get g2p test data ...
Start test g2p ...
[2024-11-14 06:55:27,892] [    INFO] - tokenizer config file saved in /home/aistudio/.paddlenlp/models/bert-base-chinese/tokenizer_config.json
[2024-11-14 06:55:27,892] [    INFO] - Special tokens file saved in /home/aistudio/.paddlenlp/models/bert-base-chinese/special_tokens_map.json
Building prefix dict from the default dictionary ...
[2024-11-14 06:55:28] [DEBUG] [__init__.py:113] Building prefix dict from the default dictionary ...
Loading model from cache /tmp/jieba.cache
[2024-11-14 06:55:28] [DEBUG] [__init__.py:132] Loading model from cache /tmp/jieba.cache
Loading model cost 0.568 seconds.
[2024-11-14 06:55:28] [DEBUG] [__init__.py:164] Loading model cost 0.568 seconds.
Prefix dict has been built successfully.
[2024-11-14 06:55:28] [DEBUG] [__init__.py:166] Prefix dict has been built successfully.
The avg WER of g2p is: 0.02401556248558565
Start sclite g2p ...
sclite: 2.10 TK Version 1.3
Begin alignment of Ref File: './exp/g2p/text.ref.clean' and Hyp File: './exp/g2p/text.g2p'
    Alignment# 9996 for speaker bak          

    Writing scoring report to './exp/g2p/text.g2p.sys'
    Writing raw scoring report to './exp/g2p/text.g2p.raw'
    Writing string alignments to './exp/g2p/text.g2p.pra'

Successful Completion