Create your own GitHub profile
Sign up for your own profile on GitHub, the best place to host code, manage projects, and build software alongside 50 million developers.
Sign up
- Nizhny Novgorod, Russia
- Sign in to view email
- https://twitter.com/dkurtaev
Pinned
545 contributions in the last year
Activity overview
Contributed to
opencv/opencv,
opencv/opencv_extra,
dkurt/cv_winter_camp_2020
and 5 other
repositories
Contribution activity
May 2020
Created a pull request in opencv/opencv that received 13 comments
Flush to zero Convolution denormal weights
model from #17259
(base) model_1: 56.99ms
(base) model_2: 170.85ms
(ftz) model_1: 49.18ms
(ftz) model_2: 49.33ms
import numpy as np
import cv2 as cv
+11
−0
•
13
comments
- Fixed NMSBoxes bug in DetectionModel
- Fixed NMSBoxes bug
- Support FlowNet2 model
- Determine SSD input shape
- Fixed bug with ngraph::op::v1::Multiply
- Added ONNX BatchNorm subgraph
- cuda4dnn(Slice): fix regression from 3.4 port
- cuda4dnn(region): add scale_x_y parameter for YOLOv4
- Prefer addAll instead of iteration for performance
- Add text recognition to examples_dnn_text_decoder
- Added NDHWC Concat support for TensorFlow