YujiaBao / Distributional-Signatures

"Few-shot Text Classification with Distributional Signatures" ICLR 2020
https://arxiv.org/abs/1908.06039
MIT License
254 stars 57 forks source link

After training? #24

Closed versipellis closed 3 years ago

versipellis commented 3 years ago

Maybe it's not super clear to me, but after you train and save the model, how do you use it?

YujiaBao commented 3 years ago

In line 294 of src/main.py, we evaluate the model on the testing data.

versipellis commented 3 years ago

Thanks. Can you not use it afterwards as a classifier on unlabeled data?

rmwu commented 3 years ago

You can directly return pred from line 70 of r2d2.py (or similar output variables in other classifiers) to obtain the model's output, pre-evaluation.

versipellis commented 3 years ago

thank you!