Paulo Valente
Paulo Valente
We don't have anyone working on Torchx currently. The next compiler we're aiming to implement is the one in EMLX, for which we have a [more concrete plan](https://github.com/elixir-nx/emlx/issues/61). However, if...
> The general idea looks neat to me! Although we still need to figure out how they would be specified via `defn`. > > We probably want to use optional...
@joshprice a PR improving the moduledoc would probably be nice. `Nx.Random.key/0` doesn't seem to be the right route to me. Take for example this code: ```elixir # Let's take key/0...
@stewartj3 I locked the conversation, but the issue will still be investigated. For some reason the issue was being targeted for spam.
I'd be happy to review/iterate on a PR over this, even if it's an opt-in interceptor or something like that. Also, consider adding telemetry events or metadata to the already...
What cudnn version are you using? I'm a bit surprised this doesn't work out of the box with the 50xx series
We have work in progress on updating XLA, which should help with this. It is mostly working, but we have to figure out how to keep supporting different CUDA versions
You are missing bazel
I suggest going to the elixir-nx/xla repository and check out the dependencies it installs
If running on livebook, ensure that the PATH is properly set. Livebook desktop will run with a different PATH configuration