Supplementry Files for MinD-Vis

This folder contains data and model checkpoints for "Seeing Beyond the Brain: Masked Modeling Conditioned Diffusion Model for Human Vision Decoding". 

There are three parts in the data folder:

1. HCP (only three dummy entries were included due to size limitations and license issues)

2. Kamitani 

3. BOLD5000

There are three subfolders in the pretrains folder:

1. ldm (downloaded from Latent Diffusion Model github repo)

2. GOD

3. BOLD5000 

fmri_encoder.pth  - SC-MBM trained on HCP + target dataset;

finetuned.pth - DC-LDM fine-tuned using pair {fMRI, Image}.