Let’s welcome Thomas Viehmann (@t-vi) as an Apache TVM reviewer!
Thomas has made valuable contributions to TVM, notably improving and maintaining the ROCm code generation backend and TOPI schedules for AMD GPUs. In addition Thomas has put together an impressive BERT training demo in Pytorch using TVM codegen by making steady enhancements to TVM including tunable schedules for batched mat mult, enhancements to the Pytorch frontend, adding a combine batch_matmul pass, and welcome improvements to Relays first order gradient graphs.