SWAFN

Code for Paper "SWAFN: Sentimental Words Aware Fusion Network for Multimodal Sentiment Analysis"

COLING2020, Pages: 1067-1077

Minping Chen, Xia Li

Requires:

python > 3.6,
pytorch == 1.4.0,
sklearn,
h5py

Datasets:

The MOSI dataset, MOSEI dataset and YouTube dataset can be downlowded at http://immortal.multicomp.cs.cmu.edu/raw_datasets/processed_data/cmu-mosi/seq_length_20/, http://immortal.multicomp.cs.cmu.edu/raw_datasets/processed_data/cmu-mosei/seq_length_20/data/ and https://github.com/pliang279/MFN/tree/master/new_data/youtube respectively. Put the downloaded datasets to the corresponding folder.
We have already provided the labels of the sentimental words classification task of the three datasets in the corresponding folder .

Run the code:

To train and test on CMU-MOSI dataset, run:
python MOSI/mosi_model.py

To train and test on CMU-MOSEI dataset, run:
python MOSEI/mosei_model.py

To train and test on YouTube dataset, run:
python YouTube/youtube_model.py