This is the official implementation for the paper "EMP: Emotion-guided Multi-modal Fusion and Contrastive Learning for Personality Traits Recognition".
- Python 3.9.1
- pytorch-lightning 1.7.2
- Linux 5.11.0-46-generic
- We use sentence transfomer for text feature extraction. Sentence Embedding
- We use large X3D network for visual features extraction. X3D
Chalearn first impressions dataset can be found in First impressions.
ELEA dataset can be found on this official website ELEA and you need to apply it.
ulimit -SHn 51200
python main.py --accelerator 'gpu' --devices 1