The source code of <<BIT: Improving Image-text Sentiment Analysis via Learning Bidirectional Image-text Interaction>> that is accepted by IJCNN 2023 (Oral).
Due to the existence of two references with the same name（same .bib in LaTex file）, an error occurred during LaTeX compilation in the final version of BIT. Therefore, the paper with the correct references is provided.
1.Processed TumEmo is (https://github.com/YangXiaocui1215/MGNNS). In our work, we use processed TumEmo following this link.
2.Original TumEmo is (https://github.com/YangXiaocui1215/MVAN).
3.MVSA-Single is (https://mcrlab.net/research/mvsa-sentiment-analysis-on-multi-view-social-data/).
1.Original clip : https://github.com/openai/CLIP
2.To get the features we need, we modified the original clip. Modified code can be found in the lines [359-372] of 'clip/model.py'.