refactor rhy

examples/other/rhy/local/pre_for_sp_aishell.py

+#!/usr/bin/env python3
 import argparse
 import os
 import re
@@ -8,7 +9,7 @@ replace_ = {"#1": "%", "#2": "`", "#3": "~", "#4": "$"}
 
 
 def replace_rhy_with_punc(line):
-    # r'[：、，；。？！,.:;"?!”’《》【】<=>{}()（）#&@“”^_|…\\]%*$', '', line)     #参考checkcheck_oov.py,
+    # r'[：、，；。？！,.:;"?!”’《》【】<=>{}()（）#&@“”^_|…\\]%*$', '', line)     #参考check_oov.py,
     line = re.sub(r'[：、，；。？！,.:;"?!’《》【】<=>{}()（）#&@“”^_|…\\]%*$', '', line)
     for r in replace_.keys():
         if r in line:

examples/other/rhy/local/pre_for_sp_csmsc.py

+#!/usr/bin/env python3
 import argparse
 import os
 import re
@@ -6,7 +7,7 @@ replace_ = {"#1": "%", "#2": "`", "#3": "~", "#4": "$"}
 
 
 def replace_rhy_with_punc(line):
-    # r'[：、，；。？！,.:;"?!”’《》【】<=>{}()（）#&@“”^_|…\\]%*$', '', line)     #参考checkcheck_oov.py,
+    # r'[：、，；。？！,.:;"?!”’《》【】<=>{}()（）#&@“”^_|…\\]%*$', '', line)     #参考check_oov.py,
     line = re.sub(r'^$\*%', '', line)
     for r in replace_.keys():
         if r in line:

examples/other/rhy/run.sh

@@ -6,9 +6,11 @@ gpus=0
 stage=0
 stop_stage=100
 
+data=data
+mkdir -p $data
+
 aishell_data=label_train-set.txt
 csmsc_data=000001-010000.txt
-processed_path=data
 
 conf_path=conf/default.yaml
 train_output_path=exp/default
@@ -23,7 +25,7 @@ source ${MAIN_ROOT}/utils/parse_options.sh || exit 1
 
 if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
     # prepare data
-    ./local/data.sh ${aishell_data} ${csmsc_data} ${processed_path}
+    ./local/data.sh ${aishell_data} ${csmsc_data} ${data}
 fi
 
 if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then

paddlespeech/text/exps/ernie_linear/train.py

@@ -66,7 +66,7 @@ def train_sp(args, config):
     seed_everything(config.seed)
 
     print(
-        f"rank: {dist.get_rank()}, pid: {os.getpid()}, parent_pid: {os.getppid()}",
+        f"rank:{dist.get_rank()}, pid: {os.getpid()}, parent_pid: {os.getppid()}"
     )
     # dataloader has been too verbose
     logging.getLogger("DataLoader").disabled = True