Yolo + Resnet50 + Resnet50 (L2) pipeline - Yolo + Resnet50 + Resnet50 (L2) pipeline - 6.3 English - XD056

Vitis Video Analytics SDK (VVAS) 6.3 User Guide (XD056)
Document_ID
XD056
Release_Date
2026-09-24
Version
6.3 English

Note

  • The following JSON examples are only samples. Ensure model-path is set to point to your available model before running.

  • Ensure label-file-path in the YOLOv5n JSON points to the correct label file for your model.

  • In branched pipelines (using tee), metadata and ROI buffers are shared by default. If attach-ppe-outbuf is enabled in one branch, its ROI buffers may be reused by other branches, which can cause classification results from different branches to appear together. To avoid cross-branch mixing, keep attach-ppe-outbuf=false for secondary stages in branched pipelines, or add a buffer/meta copy in each branch before running the second-stage vvas_xinfer.

gst-launch-1.0 \
filesrc location=<INPUT>.bgr ! \
rawvideoparse format=bgr height=<INPUT_HEIGHT> width=<INPUT_WIDTH> ! \
vvas_xinfer config-file=/etc/vvas/configs/infer/yolov5n_vart.json ! \
vvas_xinfer config-file=/etc/vvas/configs/infer/resnet50_l2.json ! \
vvas_xinfer config-file=/etc/vvas/configs/infer/resnet50_l2.json ! \
vvas_xmetaconvert config-location=/etc/vvas/configs/metaconvert/metaconvert_config.json ! \
vvas_xoverlay ! \
fpsdisplaysink sync=false text-overlay=false video-sink='filesink location=./output_yolo_r50_r18_640x640.bgr' -v