Upload main.py

main.py

@@ -0,0 +1,27 @@
+import random
+import pandas as pd
+from tqdm import tqdm
+from sklearn.model_selection import train_test_split
+
+df = pd.read_parquet('gemago2_dataset.parquet')
+
+texts = []
+for _, row in tqdm(df.iterrows(), desc="row", leave=False, total=len(df)):
+    texts.append(rf"<kor>{row['korean']}</kor>\n\n<{row['language']}>{row['target']}</{row['language']}>")
+    texts.append(rf"<{row['language']}>{row['target']}</{row['language']}>\n\n<kor>{row['korean']}</kor>")
+del df
+
+train_texts, test_texts = train_test_split(texts, test_size=0.2, random_state=42)
+del texts
+
+random.shuffle(test_texts)
+with open("test.txt", "w", encoding="UTF-8") as f:
+    f.write("\n".join(test_texts))
+del test_texts
+
+random.shuffle(train_texts)
+with open("train.txt", "w", encoding="UTF-8") as f:
+    f.write("\n".join(train_texts))
+del train_texts
+
+# gemago2_dataset_final