-
-
Notifications
You must be signed in to change notification settings - Fork 13
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.
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
introduce trilinear upsampling #5
Conversation
We should also discuss supporting interpolations.jl directly. They have added the plumbing required to make it compatible with zygote. We would still need to support getting the adjoints to the Interpolations themselves, but this is a promising move nonetheless. |
Co-authored-by: Dhairya Gandhi <dhairya@juliacomputing.com>
Fix use of convdims.
Maybe you can re-trigger the checks, as the corresponding NNlib PR has been merged. |
@CarloLucibello @DhairyaLGandhi |
Add tests for them
Update src/gather.jl Co-authored-by: Carlo Lucibello <carlo.lucibello@gmail.com> replace _check_dims as gather_check_dims fix update manifest.toml
Co-authored-by: Michael Abbott <32575566+mcabbott@users.noreply.github.com>
Add links to readme
Test `softmax` with varying dimensions
add count_indices for cuarray add CUDA kernel for divide_by_counts! add NNlib.∇scatter_src for cuda gradient support scatter mean AD for CUDA support scatter *,/ AD for CUDA
Support scatter for CUDA gradient
Use `CUDA.@atomic`
Bump patch and check for failures with Julia 1.7-nightly
remove some old cruft
Co-authored-by: Dhairya Gandhi <dhairya@juliacomputing.com>
Add group support for convolutions
I made some git chaos. Will make a new PR. |
Hi! This is the counterpart to FluxML/NNlib.jl#315 The checks fail because the NNlib PR is not yet merged.