Code submission for "Revisiting Kernel Attention via Correlated Gaussian Process Representation"
We follow the previously published implementation of the paper Calibrating Transformers via Sparse Gaussian Processes. The setup for Anaconda environment is as follows:
- Create new python environment with version python=3.8 and activate it
- Run
pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html
- Run
pip install -r requirements.txt
in this directory to install the remaining packages. - Download the CIFAR10 dataset here to data.
- Download the COLA dataset here to data.
The code for this task is in CIFAR10. Please refer to this README.md file for more details.
The code for this task is in COLA. Please refer to this README.md file for more details.