MarvellMLTools


Marvell(R) supports a family of high performance Data Processing Units (DPUs) with
integrated compute, high speed I/O and workload accelerators. These workload
accelerators include Marvell's Machine Learning Inference Processor (MLIP), a highly
optimized, integrated inference engine.


This repo hosts Marvell's ML toolchain to deploy ML models on the MLIP:

  1. bin/mrvl-tmlc is an ML compiler
  2. bin/mrvl-mlsim is a software simulator tool to simulate the MLIP hardware

This toolchain is integrated with the Apache TVM project: https://github.com/apache/tvm

Adding Marvell tools binaries to PATH will enable TVM to access the toolchain

For a detailed usage guide, please refer to: https://github.com/apache/tvm/blob/main/docs/how_to/deploy/mrvl.rst

NOTE: The toolchain is tested on an x86 platform running Ubuntu 22.04