- Python 3.8.5
- torch 1.7.1
- CUDA 11.3
The raw data can be found at IEMOCAP and MELD.
In our paper, we use pre-extracted features. The multimodal features are available at here.
For instance, to train on IEMOCAP:
python -u train.py --base-model 'LSTM' --dropout 0.5 --lr 0.00009 --batch-size 16 --model_type='emoplex' --epochs=120 --multi_modal --modals='avl' --Dataset='IEMOCAP' --norm LN