search

marcoslucianops / DeepStream-Yolo

NVIDIA DeepStream SDK 7.0 / 6.4 / 6.3 / 6.2 / 6.1.1 / 6.1 / 6.0.1 / 6.0 / 5.1 implementation for YOLO models
MIT License
1.45k stars 357 forks source link

understanding gpuYoloLayer. #233

Closed rahulsharma11 closed 1 year ago

rahulsharma11 commented 2 years ago

Hi, i want to integrate yolov5-face with landmarks.

Can you elaborate the gpuToloLayer function?

What is blockIdx.x / blockDim.x and threadIdx.x? Also how can i add landmark extraction code in gpuYoloLayer?

What is meaning of - gpuYoloLayer<<<number_of_blocks, threads_per_block, 0, stream>>>( reinterpret_cast<const float>(input) + (batch inputSize), reinterpret_cast<int>(num_detections) + (batch), reinterpret_cast<float>(detection_boxes) + (batch 4 outputSize), reinterpret_cast<float>(detection_scores) + (batch outputSize), reinterpret_cast<int>(detection_classes) + (batch outputSize), scoreThreshold, netWidth, netHeight, gridSizeX, gridSizeY, numOutputClasses, numBBoxes, scaleXY, reinterpret_cast<const float>(anchors), reinterpret_cast<const int>(mask));

Thanks.

marcoslucianops commented 2 years ago

It's not only the kernel, there are unsupported layers in the model yaml.