Added code generating the dataset

make_dataset.py

@@ -0,0 +1,21 @@
+from ast import literal_eval
+
+from datasets import Dataset, load_dataset
+
+pm = load_dataset(
+    "owaiskha9654/PubMed_MultiLabel_Text_Classification_Dataset_MeSH", split="train"
+)
+
+records = []
+for t, lab in zip(pm["abstractText"], pm["meshMajor"]):
+    try:
+        labels = literal_eval(lab)
+        labels = labels[:2]
+        records.append({"text": t, "label": labels})
+    except Exception:
+        print("Couldn't add entry:", lab)
+
+ds = Dataset.from_list(records)
+ds = ds.train_test_split(test_size=0.2, seed=42)
+
+ds.push_to_hub("kardosdrur/pubmed")