Adam Pocock

If we're talking about different packages you can use ONNX Runtime sessions one per GPU in the same JVM. But let's get it working properly in TF-Java. A while ago...

The upstream TensorFlow project doesn't yet publish aarch64 binaries on Linux, and we build all our binaries on Github Actions which doesn't provide aarch64 hosted runners, so it's difficult for...

It's saying it's a duplicate because there's a const node in your python graph and the Java code is trying to use the same name. I think the node naming...

The naming and scoping logic has been rewritten in the current `master` branch, and that will come with the `0.4.0` release. Can you check the `0.4.0-SNAPSHOT` releases to see if...

You can see a short example of loading in a saved model and executing a function from it in the TF-Java test suite (https://github.com/tensorflow/java/blob/master/tensorflow-core/tensorflow-core-api/src/test/java/org/tensorflow/SavedModelBundleTest.java#L285), and there are many more examples...

You can't do that as the `SavedModelBundle` format is a pb along with a checkpoint of the variables (well, it's conceptually a checkpoint). If you've got a frozen graph from...

We punted on this issue in Tribuo, so our sparse vectors and matrices have a fixed index structure, but the values can change. That's turned out to work quite well...

What JVM version and GC algorithm are you using?

Are you enabling any GPU specific configuration options at runtime (e.g. CUDA unified memory)? Is it possible to get a heap dump just before it runs out of memory? The...

Can you successfully use the GPU from TF 2.5.0 in Python? That will narrow down if it's a CUDA versioning issue, or something in the way we package TF Java....