jiangzhs/OpenCompass
1
0
Fork 0
mirror of https://github.com/open-compass/opencompass.git synced 2025-05-30 16:03:24 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

fix lawbench 2-1 f0.5 score calculation bug (#795)

Browse Source 
* fix lawbench 2-1 f0.5 score calculation bug

* use path in overall datasets folder

---------

Co-authored-by: Leymore <zfz-960727@163.com>
This commit is contained in:
Yggdrasill7D6 2024-02-06 22:20:11 +08:00 committed by GitHub
parent 1c8e193de8
commit a6c49f15ce
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 4 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
opencompass/datasets/lawbench/utils/modules/alignment.py
View File

@ -22,8 +22,7 @@ def read_cilin():
Cilin 詞林 is a thesaurus with semantic information
"""
# TODO -- fix this path
project_dir = os.path.dirname(os.path.dirname(__file__)) # ymliu@2023.5.30 fix the path
lines = open(os.path.join(project_dir, "data", "cilin.txt"), "r", encoding="gbk").read().strip().split("\n")
lines = open(os.path.join(os.path.dirname(__file__), "..", "..", "..", "..", "..", "data", "lawbench", "eval_assets", "cilin.txt"), "r", encoding="gbk").read().strip().split("\n")
semantic_dict = {}
semantic_classes = {}
for line in lines:
@ -40,8 +39,7 @@ def read_cilin():
def read_confusion():
confusion_dict = {}
project_dir = os.path.dirname(os.path.dirname(__file__)) # ymliu@2023.5.30 fix the path
with open(os.path.join(project_dir, "data", "confusion_dict.txt"), "r", encoding="utf-8") as f:
with open(os.path.join(os.path.dirname(__file__), "..", "..", "..", "..", "..", "data", "lawbench", "eval_assets", "confusion_dict.txt"), "r", encoding="utf-8") as f:
for line in f:
li = line.rstrip('\n').split(" ")
confusion_dict[li[0]] = li[1:]

3
opencompass/datasets/lawbench/utils/modules/classifier.py
View File

@ -10,8 +10,7 @@ Correction = namedtuple(
"inds",
],
)
file_path = os.path.dirname(os.path.abspath(__file__))
char_smi = CharFuncs(os.path.join(file_path.replace("modules", ""), 'data/char_meta.txt'))
char_smi = CharFuncs(os.path.join(os.path.dirname(__file__), "..", "..", "..", "..", "..", "data", "lawbench", "eval_assets", "char_meta.txt"))
def check_spell_error(src_span: str,
tgt_span: str,

3
opencompass/datasets/lawbench/utils/modules/tokenizer.py
View File

@ -57,8 +57,7 @@ class Tokenizer:
"""
if bpe:
from . import tokenization
project_dir = os.path.dirname(os.path.dirname(__file__))
tokenizer = tokenization.FullTokenizer(vocab_file=os.path.join(project_dir,"data","chinese_vocab.txt"), do_lower_case=False)
tokenizer = tokenization.FullTokenizer(vocab_file=os.path.join(os.path.dirname(__file__), "..", "..", "..", "..", "..", "data", "lawbench", "eval_assets", "chinese_vocab.txt"), do_lower_case=False)
results = []
for input_string in input_strings:
if not self.segmented: # 如果没有被分字，就按照每个字符隔开（不考虑英文标点的特殊处理，也不考虑BPE），否则遵循原分字结果