Pytorch Implementation of "Adaptive Co-attention Network for Named Entity Recognition in Tweets" (AAAI 2018)
(Unofficial) Pytorch Implementation of "Adaptive Co-attention Network for Named Entity Recognition in Tweets" (AAAI 2018)
$ pip3 install -r requirements.txt
Train | Dev | Test | |
---|---|---|---|
# of Data | 4,000 | 1,000 | 3,257 |
word_vector_200d.vec
) that are only in word vocab.Image features are extracted from last pooling layer of VGG16
.
If you want to extract the feature by yourself, follow as below.
data/ner_img
from original code to this repo.img_vgg_features.pt
will be saved in data
dir)$ python3 save_vgg_feature.py
Extracted features will be downloaded automatically when you run main.py
.
paper
and the original code
, so I tried to follow the paper's equations as possible.
train
, dev
, and test
dataset. (same as the original code)
Adam
optimizer instead of RMSProp
.$ python3 main.py --do_train --do_eval
F1 (%) | |
---|---|
Re-implementation | 67.17 |
Baseline (paper) | 70.69 |