This is the ShapeNet whisking dataset used in our paper:
https://arxiv.org/abs/2505.18361110hz/ is the high-variation low-fidelity dataset
1000hz/ is the low-variation high-fidelity dataset
models contains checkpoints for our two models with the best task/neural score
If you use this dataset in your work, please cite:
@misc{chung2025tactile,
title={Task-Optimized Convolutional Recurrent Networks Align with Tactile Processing in the Rodent Brain}… See the full description on the dataset page:
https://huggingface.co/datasets/NeuroAgentsLab/tactile-whisking.
