Pytorch implementation of RF_VAE proposed in Relevance Factor VAE: Learning and Identifying Disentangled Factors, Kim et al.([https://arxiv.org/abs/1902.01568])
python 3.6.4
pytorch 0.4.0 (or check pytorch-0.3.1 branch for pytorch 0.3.1)
visdom
tqdm
NHATS
ACT_Slice
NOTE: I recommend to preprocess image files(e.g. resizing) BEFORE training and avoid preprocessing on-the-fly.
initialize visdom
python -m visdom.server
e.g.
python main.py --name run_nhats --dataset nhats --gamma 6.4 --lr_VAE 1e-4 --lr_D 5e-5 --z_dim 10 ...
check training process on the visdom server
localhost:8097
##### latent traversal gif(```--save_output True```)
- Relevance Factor VAE: Learning and Identifying Disentangled Factors.([https://arxiv.org/abs/1902.01568])
- Explainable semi-supervised deep learning shows that dementia is associated with small, avocado-shaped clocks with irregularly placed hands ([https://www.nature.com/articles/s41598-023-34518-9])