use pypcre

auto_switch_cuda_toolkit.sh

@@ -79,7 +79,7 @@ config_output=$({ printf '\n'; } | ${SUDO} update-alternatives --config cuda 2>&
 
 selection=$(CONFIG_OUTPUT="${config_output}" "${python_cmd}" - "${target_version}" <<'PY'
 import os
-import re
+import pcre as re
 import sys
 
 target = sys.argv[1]

gptqmodel/adapter/adapter.py

@@ -4,7 +4,7 @@
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
 
-import re
+import pcre as re
 from dataclasses import dataclass
 from typing import Dict, List, Optional, Tuple, Union
 

gptqmodel/models/writer.py

@@ -9,7 +9,7 @@
 import csv
 import json
 import os
-import re
+import pcre as re
 from os.path import isfile, join
 from typing import Any, Dict, Optional, Union
 

gptqmodel/quantization/config.py

@@ -5,7 +5,7 @@
 
 import json
 import os.path
-import re
+import pcre as re
 from dataclasses import dataclass, field, fields
 from enum import Enum
 from os.path import join

gptqmodel/utils/mmlupro.py

@@ -7,7 +7,7 @@
 import json
 import os
 import random
-import re
+import pcre as re
 import time
 
 import torch

gptqmodel/utils/model.py

@@ -11,7 +11,7 @@
 import math
 import operator
 import os
-import re
+import pcre as re
 import shutil
 import struct
 import threading

gptqmodel/utils/structure.py

@@ -24,7 +24,7 @@
 - Collapsing is generic: any numeric-indexed ModuleList whose qualified name matches `experts-regex`.
 """
 
-import re
+import pcre as re
 from typing import Dict, Iterable, Optional, Set, Tuple
 
 import torch

pyproject.toml

@@ -41,6 +41,7 @@ dependencies = [
     "device-smi>=0.5.1",
     "protobuf>=6.32.0",
     "pillow>=11.3.0",
+    "pypcre>=0.2.1",
     "hf_transfer>=0.1.9",
     "huggingface_hub>=0.34.4",
     "random_word>=1.0.13",

requirements.txt

@@ -8,6 +8,7 @@ packaging>=24.2
 device-smi>=0.5.1
 protobuf>=6.32.0
 pillow>=11.3.0
+pypcre>=0.2.1
 hf_transfer>=0.1.9
 huggingface_hub>=0.34.4
 random_word>=1.0.13

setup.py

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 import os
-import re
+import pcre as re
 import subprocess
 import sys
 from pathlib import Path

tests/models/test_llama3_2.py

@@ -23,7 +23,7 @@ class TestLlama3_2(ModelTest):
     DATASET_SIZE = 1024
     DATASET_SORT = "desc"
     QUANT_BATCH_SIZE = 4
-    # USE_FLASH_ATTN = False
+    USE_FLASH_ATTN = True
     # EORA = Lora(
     #     # for quant, path is save path. for load, it is loading path
     #     path="./eora_test",

tests/tasks/gpqa/cot_n_shot/utils.py

@@ -4,7 +4,7 @@
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
 import random
-import re
+import pcre as re
 
 import datasets
 

tests/tasks/gpqa/cot_zeroshot/utils.py

@@ -4,7 +4,7 @@
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
 import random
-import re
+import pcre as re
 
 import datasets
 

tests/tasks/gpqa/generative/utils.py

@@ -4,7 +4,7 @@
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
 import random
-import re
+import pcre as re
 
 import datasets
 

tests/tasks/gpqa/n_shot/utils.py

@@ -4,7 +4,7 @@
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
 import random
-import re
+import pcre as re
 
 import datasets
 

tests/tasks/gpqa/zeroshot/utils.py

@@ -4,7 +4,7 @@
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
 import random
-import re
+import pcre as re
 
 import datasets
 

tests/tasks/hellaswag/utils.py

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
-import re
+import pcre as re
 
 import datasets
 

tests/tasks/mmlu/flan_cot_zeroshot/utils.py

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
-import re
+import pcre as re
 import sys
 import unicodedata
 

tests/tasks/mmlu/flan_n_shot/generative/utils.py

@@ -3,7 +3,7 @@
 # SPDX-License-Identifier: Apache-2.0
 # Contact: qubitium@modelcloud.ai, x.com/qubitium
 
-import re
+import pcre as re
 import sys
 import unicodedata