I followed the dataflow compiler tutorial and jupyter notebooks on my main rig, those examples worked. I used them to take this model: https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5m.onnx to a hef file for the hailo 8l.
I am able to run on the rpi5:
matteius@matteius-desktop:~/Projects/tappas/apps/h8/gstreamer/raspberrypi/detection$ hailortcli run ~/yolov5m.hef
Running streaming inference (/home/matteius/yolov5m.hef):
Transform data: true
Type: auto
Quantized: true
Network yolov5m/yolov5m: 100% | 81 | FPS: 16.20 | ETA: 00:00:00
> Inference result:
Network group: yolov5m
Frames count: 81
FPS: 16.20
Send Rate: 159.22 Mbit/s
Recv Rate: 278.63 Mbit/s
But my question is how do I actually utilize the model? I tired modifying the detection.sh example:
diff --git a/apps/h8/gstreamer/raspberrypi/detection/detection.sh b/apps/h8/gstreamer/raspberrypi/detection/detection.sh
index 1358cf9..1a3e452 100755
--- a/apps/h8/gstreamer/raspberrypi/detection/detection.sh
+++ b/apps/h8/gstreamer/raspberrypi/detection/detection.sh
@@ -74,7 +74,11 @@ function parse_args() {
hef_path="$RESOURCES_DIR/ssd_mobilenet_v1.hef"
postprocess_so="$POSTPROCESS_DIR/libmobilenet_ssd_post.so"
thresholds_str=""
- elif [ $2 != "yolov5" ]; then
+ elif [ $2 == "yolov5" ]; then
+ network_name="hailo_yolo_inference"
+ hef_path="/usr/share/hailo-models/yolov5m.hef"
+ postprocess_so="/usr/local/lib/aarch64-linux-gnu/rpicam-apps-postproc/libyolo_hailortpp_post.so"
+ else
echo "Received invalid network: $2. See expected arguments below:"
print_usage
exit 1
matteius@matteius-desktop:~/Projects/tappas/apps/h8/gstreamer/raspberrypi/detection$ ./detection.sh --network yolov5
Running hailo_yolo_inference
gst-launch-1.0 filesrc location=/home/matteius/Projects/tappas/apps/h8/gstreamer/raspberrypi/detection/resources/detection.mp4 name=src_0 ! qtdemux ! h264parse ! avdec_h264 max_threads=2 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! videoscale n-threads=2 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! videoconvert n-threads=3 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! hailonet hef-path=/usr/share/hailo-models/yolov5m.hef batch-size=1 output-format-type=HAILO_FORMAT_TYPE_FLOAT32 nms-score-threshold=0.3 nms-iou-threshold=0.45 output-format-type=HAILO_FORMAT_TYPE_FLOAT32 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! hailofilter function-name=hailo_yolo_inference so-path=/usr/local/lib/aarch64-linux-gnu/rpicam-apps-postproc/libyolo_hailortpp_post.so qos=false ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! hailooverlay ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! videoconvert n-threads=3 ! fpsdisplaysink video-sink=ximagesink name=hailo_display sync=false text-overlay=false
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Redistribute latency...
Redistribute latency...
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 131 (XInputExtension)
Minor opcode of failed request: 46 ()
Value in failed request: 0xd
Serial number of failed request: 61
Current serial number in output stream: 65
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
NMS score threshold is set, but there is no NMS output in this model.
CHECK_SUCCESS failed with status=6
Then when I take that gstreamer command and remove the nms output args I get:
matteius@matteius-desktop:~/Projects/tappas/apps/h8/gstreamer/raspberrypi/detection$ ./detection.sh --network yolov5
Running hailo_yolo_inference
gst-launch-1.0 filesrc location=/home/matteius/Projects/tappas/apps/h8/gstreamer/raspberrypi/detection/resources/detection.mp4 name=src_0 ! qtdemux ! h264parse ! avdec_h264 max_threads=2 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! videoscale n-threads=2 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! videoconvert n-threads=3 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! hailonet hef-path=/usr/share/hailo-models/yolov5m.hef batch-size=1 output-format-type=HAILO_FORMAT_TYPE_FLOAT32 output-format-type=HAILO_FORMAT_TYPE_FLOAT32 ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! hailofilter function-name=hailo_yolo_inference so-path=/usr/local/lib/aarch64-linux-gnu/rpicam-apps-postproc/libyolo_hailortpp_post.so qos=false ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! hailooverlay ! queue max-size-buffers=5 max-size-bytes=0 max-size-time=0 ! videoconvert n-threads=3 ! fpsdisplaysink video-sink=ximagesink name=hailo_display sync=false text-overlay=false
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Redistribute latency...
Redistribute latency...
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 131 (XInputExtension)
Minor opcode of failed request: 46 ()
Value in failed request: 0xd
Serial number of failed request: 61
Current serial number in output stream: 65
