Yufei Wu

[Suggestion] Improve Pedestrain Trackor C++ Demo

6

Hi, OpenVINO guys, I learn a lot from your Pedestrain Trackor demo, however I wonder why some intel model can't be supported in this demo. Such as person-detection-0203, person-detection-0106. Only...

How to use latest OpenVINO with dl streamer？

11

Is there anyway we can use the self-built openvino with dl streamer without using the APT install one？

gvatrack-GPU performance drops comparing with CPU implementation

9

Hi, I want to check whether GVATrack GPU implementation could benefit my project, so I evalution the performance in devcloud-oneAPI: `qsub -l nodes=s012-n002:gpu:ppn=2:iris_xe_max -d . tracker.sh` The performance shows GPU-gavtrack...

Usgae: run ./human_pose_estimation.sh [INPUT_FILE] postprocessing codes for heatmap are copied from open model zoo

Real-time performance isn't so good

2

Hi, I have tried the inference code in NYU dataset. However, I can't achieve "real-time" performance as mentioned in your paper. For batch size=1: frame rate is 12 For batch...

1. hourglass module should have symmetric structure which means it has three residual modules in at the most inner layer. ``` if n > 1: lower2 = _hourglass(lower1, n-1) else:...

Update smart science lab demo

3

Update UI and modify some demo logic including mstcn, skip frame.

mobilenet-v3 without classifier layer -> smartlab-sequence-modelling-0001 mstcn++ -> smartlab-sequence-modelling-0002 Object detection models may have updated accuracy, wait for feedback now.

Lack of except check in Tensor.py, which will lead to runtime error. If raw inference data can't read, just return None.

call_origin means such function can't executed in GPU?

1

Hi, I want to use dpnp.convolve, but the source code: ` return call_origin(numpy.convolve, a=a, v=v, mode=mode) ` So, I wonder whether this function can be execued in GPU or not?...