HirataYurina

yolov4-tiny-keras

yolov4-tiny is a good choice when you don't have expensive devices.

simple-pose-estimation-torch-techi

Only 200 lines code to predict single person pose estimation. A simple prediction module of "Simple Baselines for Human Pose Estimation and Tracking".

yoloV4-keras-techi

yoloV4 implemented by keras and tf backend. Enjoy it, my homie.

RetinaNet-tf2.X-sibyl

use eager tf to do my projects because it is difficult to debug with frozen graph

yoloV3-keras-sibyl

so cool. I used yolov3, gaussian yolov3, asff yolov3 to do my project.

ShuffleNetV2-SqueezeNet-Inception

ShuffleNetV2, SqueezeNet, inceptionV1 implemented by tensorflow2.0

sort-YOLOv4-techi

My yolo4 and sort codes that implement avoiding invasion module.

weighted-boxes-fusion

Weighted Boxes Fusion

DarkFantasy

computer vision kit implemented by tf2.0

cascade-rcnn-tf2.2

cascade rcnn, faster rcnn implemented by tensorflow2.2 and GHM CLoss, GHM RLoss, Pseudo Labeling

machine-learning

Just Machine Learning and I think Machine Learning is more difficult than Deep Learning.

torch-onnx-tensorflow

use onnx inference architecture in intelligent construction. (.pth to .onnx to .pb)

data-augment-toolkit-numpy

a data augment toolkit

frequently-used-website

life is a struggle

model-metric-techi

use mAP, recall, precision, f-measure, roc, auc to estimate your model.

tensorflow-deploy

use tensorflow-serving to deploy projects, use tensorRT to accelerate projects, exchange FP32 to FP16, FP32 to INT8

danger_source_crawler

EfficientDet-tensorflow2

EfficientDet implemented by tensorflow2.0

insightface-retinaface-tf-techi

Just use a little codes to implement insight face and retina face. And i apply this algorithm into intelligent construction.

Mask-Rcnn-tf2

use tensorflow2 to implement mask rcnn

DDPG-tensorflow2.2

DDPG algorithm written by tensorflow2.2

DuelingDeepQLearning

reinforcement learning

kaggle-lirui

some kaggle code

OpenCL-CUDA-Performance

SVM-C-plus

C++ codes

clpeak

A tool which profiles OpenCL devices to find their peak capacities

clspv

Clspv is a prototype compiler for a subset of OpenCL C to Vulkan compute shaders

mixbench

A GPU benchmark tool for evaluating GPUs on mixed operational intensity kernels (CUDA, OpenCL, HIP, SYCL)

thrift

Apache Thrift

tvm

Open deep learning compiler stack for cpu, gpu and specialized accelerators