Problem with NvDCF tracker on deepstream 5.0 and jetson

AI & Data Science Deep Learning (Training & Inference)
#1

Hello,

I’m trying to run NvDCF tracker on a deepstream app on jetson AGX, but the app suddenly stops in the beginning showing the below errors and warnings.

Screenshot from 2020-06-01 10-45-36
Screenshot from 2020-06-01 10-45-36901×125 33.2 KB

Screenshoht from 2020-06-01 10-46-13
Screenshoht from 2020-06-01 10-46-131058×159 35 KB

I’m using yolo3-tiny as a detector and this is the config file of the tracker :

[tracker]
tracker-width=960
tracker-height=544
gpu-id=0
ll-lib-file=/opt/nvidia/deepstream/deepstream-5.0/lib/libnvds_nvdcf.so
ll-config-file=tracker_config.yml
enable-batch-process=1

Also this is the app config file:

[property]
gpu-id=0
interval=1
gie-unique-id=1
batch-size=1
nvbuf-memory-type=1
net-scale-factor=0.0039215697906911373
#0=RGB, 1=BGR
model-color-format=0
network-type=0
config-file=/opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/config_infer_primary_yoloV3_tiny.txt
custom-network-config=yolov3-tiny_RL.cfg
model-file=yolov3-tiny_RL.weights
model-engine-file=model_b1_gpu0_int8.engine
labelfile-path=labels.txt
int8-calib-file=/opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/yolov3-calibration.table.trt7.0
network-mode=1
num-detected-classes=1
is-classifier=0
maintain-aspect-ratio=1
parse-bbox-func-name=NvDsInferParseCustomYoloV3Tiny
engine-create-func-name=NvDsInferYoloCudaEngineGet
custom-lib-path=/opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/nvdsinfer_custom_impl_Yolo/libnvdsinfer_custom_impl_Yolo.so
process-mode=1
#0=Group Rectangles, 1=DBSCAN, 2=NMS, 3 = None(No clustering)
cluster-mode=3

[class-attrs-all]
nms-iou-threshold=0.45
threshold=0.5

#2

I commented minMatchingScore4Motion and matchingScoreWeight4Motion in tracker_config.yml, the error has changed to:

Screenshot from 2020-06-07 11-49-13
Screenshot from 2020-06-07 11-49-13733×119 11.1 KB

However, I tried to run source12_1080p_dec_infer-resnet_tracker_tiled_display_fp16_tx2.txt example with NvDCF tracker and I worked. I’m not really sure what caused the problem. I also tried to run my app without tracker_config.yml and it worked. So, anyone know what is the reason behind it? Also is the tracker very slow or there is something wrong with it?

I’m using Python version of deepstream.

#3

I manage to solve the problem by commenting endFrameNum: 200 line in tracker_config.yml.

2 Likes