Collective Knowledge repository for collaboratively benchmarking and optimising embedded deep vision runtime library for Jetson TX1

Introduction

CK-TensorRT is an open framework for collaborative and reproducible optimisation of convolutional neural networks for Jetson TX1. It's based on the Deep Inference framework from the Dustin Franklin (Jetson Developer @NVIDIA) and the Collective Knowledge framework from the cTuning Foundation. In essence, CK-TensorRT is simply a suite of convenient wrappers for building, evaluating and optimising performance of Jetson Inference runtime library for Jetson TX1.

Authors/contributors

• Anton Lokhmotov, dividiti
• Daniil Efremov, xored

Quick installation on Ubuntu

TBD

Installing general dependencies

$ sudo apt install coreutils \
                   build-essential \
                   make \
                   cmake \
                   wget \
                   git \
                   python \
                   python-pip

Installing CK-TensorRT dependencies

TBD

Installing CK

$ sudo pip install ck
$ ck version

Installing CK-TensorRT repository

$ ck pull repo:ck-tensorrt --url=https://github.com/dividiti/ck-tensorrt

Building CK-TensorRT and all dependencies via CK

The first time you run a TensorRT program (e.g. tensorrt-test), CK will build and install all missing dependencies on your machine, download the required data sets and start the benchmark:

$ ck run program:tensorrt-test

Related projects and initiatives

We are working with the community to unify and crowdsource performance analysis and tuning of various DNN frameworks (or any realistic workload) using Collective Knowledge Technology:

• CK-Caffe
• CK-TinyDNN
• Android app for DNN crowd-benchmarking and crowd-tuning
• CK-powered ARM workload automation

Open R&D challenges

We use crowd-benchmarking and crowd-tuning of such realistic workloads across diverse hardware for open academic and industrial R&D challenges - join this community effort!

Related publications with long term vision

• All references with BibTex