-
-
Notifications
You must be signed in to change notification settings - Fork 151
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
GPU broadcast resulted in non-concrete element type Any when running several NODEs in a single training #852
Comments
|
Running on GPUs doesn't allow the use of I'm basing myself on the case with a single NODE that's on the tutorial. I'm pretty sure there's a simpler way to predict several NODEs with different initial conditions and time steps (maybe something with multiple broadcasts?), but it escapes my head |
Partitioned the predict function a little bit, still getting the same error.
|
|
Using
|
Yes that is expected. See my previous reply
|
Makes sense. What would be an appropriate implementation? Based on the documentation of LuxDeviceUtils, I tried changing Lux.gpu to gpu_device, but it didn't work.
and got
|
Even |
In that case, would |
Ok I got the syntax of
|
I'm running the following MWE:
but it leads to:
This error doesn't show up when running
loss_function(p)
, so that means that there is something going on in the optimization process. This error doesn't show up when training on CPU either.The text was updated successfully, but these errors were encountered: