albanD
albanD
cc @datumbox @mthrok what do you guys think about this?
Rendered version https://github.com/ROCmSoftwarePlatform/rfcs/blob/rfc-unified-memory/RFC-0023-Unified-Memory-for-Pytorch.md.
That's interesting... Could you share a code sample that we can use to reproduce that (on colab for example)?
Hi, It is not very common to have big enough models for this to happen no. I'm sure we could refactor the current recursive algorithm into an iterative one though...
My guess, based on the stack you shared, is just that the graph we're trying to build has too much depth. And because we use recursion every time to get...
The CUDA builds are not mentioned in this doc. Does it need any special casing?
libtorch failure is due to infra issues, you can ignore.
cc @rohan-varma who helped me with build issues on gloo. Maybe this one as well is already fixed in latest gloo and we need to update the submodule?
ping @mrshenli
Hey, There are no custom shaders at the moment as everything we needed for the basic networks we looked at was already provided by MPS (or a set of ops...