Datasets:
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") | |