AMontgomerie/question_generator

missing dataset

Opened this issue · 3 comments

Deleted user commented

Hi @iarfmoose
Thanks for sharing this T5 example on question generation task.
The dataset you used for training is not included in the repo

train_set = QGDataset(os.path.join(DIR, 'question_generator/datasets/qg_train.csv')) train_loader = DataLoader(train_set, batch_size=BATCH_SIZE, shuffle=True) valid_set = QGDataset(os.path.join(DIR, 'question_generator/datasets/qg_valid.csv')) valid_loader = DataLoader(valid_set, batch_size=BATCH_SIZE, shuffle=False)

Would you be so kind to provide them ?
Thanks in advance
Philippe

AMontgomerie commented

Hi Philippe,

The datasets are too large to upload to the repo but here's a google drive download link

Deleted user commented

Thanks a lot @iarfmoose
Have you found this repo already: https://github.com/patil-suraj/question_generation ?
Cheers

AMontgomerie commented

No problem!

Yeah I came across that one after uploading my project. It seems to be a lot more in-depth than mine!