Abstract: The past few years, traditional compiler optimization methods have been found to be further enhanced by machine learning (ML), deep learning (DL) and reinforcement learning (RL). These ...

I have generated a PB (Protocol Buffer) file or an MHLO file or an stablehlo file from tensorflow. How can I perform AOT compilation and deployment in xla using these files? From the source code and ...

First of all, thank you for the new AOT feature. I’m testing some kernels using CuTeDSL AOT with TVM-FFI, and I’m seeing a significant reduction in CPU overhead. I’d like to compile/use many functions ...