Apache TVM VTA hardware components

Clone this repo:

Branches

  1. 21937a0 [pynq_driver] fix device early return (#7) by ZHANG Hao · 5 weeks ago master
  2. 22e4b88 [Hardware][Xilinx] utilize uram (#6) by ZHANG Hao · 5 weeks ago
  3. db65157 [xilinx] correct assert statement (#4) by ga · 10 weeks ago
  4. 6beda21 [CI] Add vta path to cpptest (#3) by Tianqi Chen · 10 weeks ago
  5. 6905fc3 Update ASF yaml (#2) by Tianqi Chen · 10 weeks ago

VTA Hardware Design Stack

Build Status

VTA (versatile tensor accelerator) is an open-source deep learning accelerator complemented with an end-to-end TVM-based compiler stack.

The key features of VTA include:

  • Generic, modular, open-source hardware
    • Streamlined workflow to deploy to FPGAs.
    • Simulator support to prototype compilation passes on regular workstations.
  • Driver and JIT runtime for both simulator and FPGA hardware back-end.
  • End-to-end TVM stack integration
    • Direct optimization and deployment of models from deep learning frameworks via TVM.
    • Customized and extensible TVM compiler back-end.
    • Flexible RPC support to ease deployment, and program FPGAs with the convenience of Python.