New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

High Level Features #186

Open

hypnopump wants to merge 1 commit into getkeops:main from hypnopump:high-level

hypnopump commented Jul 31, 2021

This PR implements some high level features such as:

Interface to call torch.nn.Linear() the kernel way.
Reproducible conda environment for pykeops in Linux-gpu
Naive implementation for several activation functions
Simple MH Attention Layer (probably to move as an example?)

Projected roadmap:

Provide interface for simple MLPs
Optimized activation functions (to do once the engine is rewritten)

Proposals:

Element-wise maximum and minimum operator for multiple tensors : max([0,1], [1,0]) = [1,1]


          initial add files

d3409e1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet