Open, Modular, Deep Learning Accelerator

Clone this repo:


  1. 1454c96 Update CI badge location (#12) by Tianqi Chen · 9 days ago master
  2. 87ce9ac [Vivado] Export VTA configuration variables to TCL for hardware builds (#11) by Luis Vega · 12 days ago
  3. 9b178fd fixes for vivado 2020.1 (#10) by Luis Vega · 4 weeks ago
  4. f1c3385 Add c++ and python local deploy example (#5) by Hua Jiang · 4 months ago
  5. 21937a0 [pynq_driver] fix device early return (#7) by ZHANG Hao · 5 months 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.