qhd0081

How to reproduce the behaviour --------- Your Environment --------- * Operating System: * Python Version: * Package Version: seq = ['O', 'B-PER', 'I-LOC', 'O', 'B-LOC', 'I-PER'] print(get_entities(seq)) [('PER', 1, 1),...

I wonder, to extract all types of eneities, does the model asks several questions to extract different types of entities or just ask one question a time and do this...

How to use the model with sentence-transformer for inference?

2

Cannot load the model. code from sentence_transformers import SentenceTransformer model = SentenceTransformer("../../models/consbert/unsup-consert-base-atec_ccks") # the model path Error message Traceback (most recent call last): File "/home/qhd/PythonProjects/GraduationProject/code/preprocess_unlabeled_second/sentence-bert.py", line 16, in model =...

Why the corpus is already tokenized？

1

Why there are many blamks in the text? I mean the text is already tokenized.

why still need keras_bert?

1

I have run this code and it asks me to install keras_bert. Thoutht it would be a pytorch reimplementation as the repo name says.

@ZeroRin Just averaging the loss of in one batche may be confusing for the following reasons: 1) Since the number of training samples in one batch are not always the...

Would you mind providing the exact F1 and EM of each domain for future comparison? The results in the paper are showned in bar graph and the exact F1 and...