varsha33 / LCL_loss

Label-aware Contrastive Loss

This repo contains the pytorch implementation of Label-aware Contrastive Loss (LCL).

Dataset preprocess

Download datasets from the given data source links, Empathetic Dialogue, GoEmotions, ISEAR, EmoInt SST-2,SST-5

For Empathetic Dialogue dataset, an additional extraction of the raw data was done to get the csv files using ed_data_extract.py which is required for the below pre-processing

python data_preprocess.py -d <dataset name> --aug ## for emotion datasets
python sst_data_preprocess.py -d <dataset name> --aug ## for sentiment datasets

Training the model

Set the parameters in config.py

python train.py

Credits

We took help from the following open source projects and we acknowledge their contributions.

1. Supervised Contrastive Loss https://github.com/HobbitLong/SupContrast
2. SST-tree2tabular https://github.com/prrao87/fine-grained-sentiment/blob/master/data/sst/tree2tabular.py
3. Tweet preprocess https://github.com/cbaziotis/ntua-slp-semeval2018.git
4. Tweet preprocess https://github.com/abdulfatir/twitter-sentiment-analysis.git

Note: There is minor typo in Eq 3 in the paper relating to the placement of weight wi yi (i.e the positive weights which are constant for a given sample). It is meant to be outside the log and is correctly implemented in the code. The results in the paper follow the current implementation.