fix(bug).when generating text that contains only punctuation marks or whitespace characters, the CPU usage reaches 100%, and the process crashes.

cosyvoice/utils/common.py

@@ -20,7 +20,6 @@ from typing import List
 
 import numpy as np
 import torch
-import regex
 
 IGNORE_ID = -1
 
@@ -156,12 +155,6 @@ def set_all_random_seed(seed):
     torch.cuda.manual_seed_all(seed)
 
 
-def is_only_punctuation(text):
-    # Regular expression: Match strings that consist only of punctuation marks or are empty.
-    punctuation_pattern = r'^[\p{P}\p{S}]*$'
-    return bool(regex.fullmatch(punctuation_pattern, text))
-
-
 def mask_to_bias(mask: torch.Tensor, dtype: torch.dtype) -> torch.Tensor:
     assert mask.dtype == torch.bool
     assert dtype in [torch.float32, torch.bfloat16, torch.float16]

cosyvoice/utils/frontend_utils.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 
 import re
+import regex
 chinese_char_pattern = re.compile(r'[\u4e00-\u9fff]+')
 
 
@@ -127,3 +128,9 @@ def replace_blank(text: str):
         else:
             out_str.append(c)
     return "".join(out_str)
+
+
+def is_only_punctuation(text):
+    # Regular expression: Match strings that consist only of punctuation marks or are empty.
+    punctuation_pattern = r'^[\p{P}\p{S}]*$'
+    return bool(regex.fullmatch(punctuation_pattern, text))