Skip to content
GammaLab
C++ CMake
Branch: master
Clone or download
Latest commit 8633af4 Jul 4, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
model ImageNet Inference V2 Jul 1, 2019
CMakeLists.txt origin commit Jun 18, 2019
CalibrationTableresnet50_4 ImageNet Inference V2 Jul 1, 2019
GammaContext.h change pre_process to pre_process_gpu Jun 23, 2019
README.md Update README.md Jul 4, 2019
gt.txt origin commit Jun 18, 2019
libGammaContext.so ImageNet Inference V2 Jul 1, 2019
test.cpp ImageNet Inference V2 Jul 1, 2019

README.md

Performance

latency without preprocess: 0.82ms

latency with preprocess: 1.26ms

top5 error: 0.0699

Enviroment Prepare

Host: 1 nvidia Tesla T4 , 1 Xeon(G) Gold 6130 cpu , 128G cpu memory

Software: ubuntu 16.04 , nvidia driver 410.104 , cuda 10.0 , cudnn 7.5 , tensorrt 5.1.5.0

Model: caffe model resnet50

Usage

Modify 'model_path, imgs_path, gt_path' in test.cpp.

mkdir build && cd build && cmake ..
make
./AIProject
You can’t perform that action at this time.