MER Experiment Notebooks: LSTM and Cross-Modal Transformer Experiments (CMU-MOSEI)

This collection contains the two experimental Jupyter notebooks for DF2025-069, implementing and evaluating three multimodal emotion recognition architectures on CMU-MOSEI.Notebook 1 (MER-Base, MER-Attn): LSTM-based fusion experiments using OpenFace 2.0 visual features and CrossEntropyLoss.Notebook 2 (MER-CMT): Cross-Modal Transformer experiments using FACET 4.2 visual features and BCEWithLogitsLoss.Full paper: https://doi.org/10.6084/m9.figshare.31567024Repository: https://github.com/visionbyangelic/Multimodal-Emotion-RecognitionSupervisor: Winner Emeto (DataraFlow) | dataraflow@gmail.com​​​​​​​​​​​​​​​​