Mingsheng Hong

Can you please provide some examples / use cases of "MLIR compiled functions"? TFRT supports some form of JIT (implemented for Google TPU, but not open sourced). Please share your...

TFRT will support different flavors of kernels: codegen'ed (via xla, mlir, or other technology), hand-written (e.g. via Eigen), and library-based (e.g. by calling into cuDNN). codengen'ed kernels can be AOT...

When calling existing kernels, indeed some overhead will be incurred, but not all kernel invocation will be on the critical path of the model computation. Our goal in that case...

We will lazily write native kernels based on performance needs. There is probably a skewed distributions of the kernels, such that we need not rewrite thousands of them. For the...

I'm not sure how you drew the connection. Can you elaborate on your reasoning and/or question?

I know this may not be saying much: Independent of the "skewed distributions" observation, TF will continue to improve its kernel technology, including fusion passes.

Will share the info down the road when ready. Thanks.

Such kernel fusion scenarios will be supported (design is WIP). I don't think this has to be done by introducing a new op to the graph though; it could be...

It's WIP and will be open sourced down the road when we are ready. Thanks for your interest.

The supplied build.xml file can be used to build this counting-triangle project and run it. That includes the support of configuring and running Hadoop on the local node. See README...