Andrew Luo

https://u.pcloud.link/publink/show?code=XZhLWj5Zd01s27dcrQ870TbFptYiNRUMWj57 Please find the weights here. In general, we use a linear layer applied to the unit normalized (OpenCLIP) CLIP ViT-B/16 output. If you use ridge regression, I recommend that...

Due to the change in the diffusers' library API, you may have to use a version from pre-2023 May.

Perhaps it is better if you reach out to me directly, you can find my contact on https://andrewluo.net

``` import os import numpy as np import nibabel as nib def load_from_nii(nii_file): return nib.load(nii_file).get_fdata() import cortex import matplotlib.pyplot as plt nsd_root = '/lab_data/tarrlab/common/datasets/NSD' for ss in [1,2,3,4,5,6,7,8]: print(ss) for...

[food_voxels.zip](https://github.com/user-attachments/files/20856843/food_voxels.zip) And here are the food voxel masks