thuiar/MMSA
MMSA is a unified framework for Multimodal Sentiment Analysis.
1 Releases
Latest: 3y ago
v_2.2.1Latest
📋 Changes
- Add `MMSA_test()` which runs MSA inference on single videos.
- Add modality missing tasks.
- Add option `custom_feature` for model training.
- Add models `TFR_NET`, `MMIM`
