Skip to content

fix: 각 converter들이 모두 같은 noised sample을 사용할 수 있도록 수정 #61

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: 각 converter들이 모두 같은 noised sample을 사용할 수 있도록 수정 #61

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 6 additions & 5 deletions ConvertedTester.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,7 @@ def _calcuate_accuracy(self, original_sentences: List[str], converted_sentences:
correct_count += 1
return correct_count / len(original_sentences)

def test(self, df: pd.DataFrame) -> Tuple[float, float, float]:
def test(self, original_df: pd.DataFrame, noised_df: pd.DataFrame) -> Tuple[float, float, float]:
"""
주어진 문장들에 대해 변환기의 성능을 테스트하는 메소드

Expand All @@ -66,8 +66,8 @@ def test(self, df: pd.DataFrame) -> Tuple[float, float, float]:
Returns:
float: 변환기의 성능
"""
sentences = df['text'].tolist()
noised_sentences_df = NoiseGeneratorASCII(ratio = 0.3).generate(df)
sentences = original_df['text'].tolist()
noised_sentences_df = noised_df.copy()
converted_sentences = self.converter.convert(noised_sentences_df)['text'].tolist()

for i, (original, noised ,converted) in enumerate(zip(sentences, noised_sentences_df['text'].to_list(),converted_sentences)):
Expand All @@ -88,10 +88,11 @@ def test(self, df: pd.DataFrame) -> Tuple[float, float, float]:
from src.data_control.noise_converter.NCVarco import NCVarco

model_ids = ['rtzr/ko-gemma-2-9b-it', 'HumanF-MarkrAI/Gukbap-Gemma2-9B', 'NCSOFT/Llama-VARCO-8B-Instruct']
prompts = ['prompts/gemma-converter.json', 'prompts/gukbap-gemma-converter.json', 'varco-converter.json']
prompts = ['prompts/gemma-converter.json', 'prompts/gukbap-gemma-converter.json', 'prompts/varco-converter.json']

sample_df = pd.read_csv('data/cleaned.csv')
sample_df = sample_df[sample_df['noise'] == 0].sample(10, random_state=42)
noised_sample_df = NoiseGeneratorASCII(ratio=0.5).generate(sample_df)

for i, model_id in enumerate(model_ids):
if 'gemma' in model_id.lower():
Expand All @@ -102,4 +103,4 @@ def test(self, df: pd.DataFrame) -> Tuple[float, float, float]:
converter = NCVarco(model_id, prompt= prompts[i])

tester = ConvertedTester(converter)
tester.test(sample_df)
tester.test(sample_df, noised_sample_df)