Model not working after upgrading from TensorRT 8.6 to TensorRT10.3 (original) (raw)

Description

A clear and concise description of the bug or issue.

Environment

TensorRT Version: 10.3.0
GPU Type: Jetson Orin AGX
Nvidia Driver Version:
CUDA Version: 12.6
CUDNN Version: 9.3
Operating System + Version: Ubuntu 22.04
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag): Baremetal

Hi,
I was running a person detect YOLOv8 model as PGIE and person attribute model as SGIE in deepstream 7.0, jetpack 6.0 and tensorRT 8.6. Now I have upgraded it as per configuration mentioned above. I have build the new engine files for both models using deepstream7.1. I found that YOLOV8 model is working properly, however the person attribute model is not able to recognize any attribute.
To debug I print confidence in custom parser and found that all the values are less than 0.2 while earlier with 7.0 I was getting values higher then 0.55.

Relevant Files

model files are available at

solider_model.onnx_b1_gpu0_fp32.engine - tensorRT 10.3 and deepstream 7.1, model was generated automatically by deepstream-app (This model inference has very low confidence score)

sim_solider.onnx_b1_gpu0_fp32.engine - tensorRT 8.6 and deepstream7.0, model was automatically generated by deepstream-app (This model working properly)

shared library and yolo model is also in this folder.

Steps To Reproduce

deepstream-app -c main_config.txt

Below is the configuration I am using.

solider_config.txt

[property]
gpu-id=0
onnx-file=solider_model.onnx
model-engine-file=solider_model.onnx_b1_gpu0_fp32.engine
batch-size=1
net-scale-factor=0.00392156862745098
offsets=123.675;116.28;103.53
model-color-format=0
network-mode=0
process-mode=2
network-type=1
labelfile-path=labels.txt
interval=0
infer-dims=3;256;192
maintain-aspect-ratio=1
symmetric-padding=0
custom-lib-path=parser/libnvdsinfer_custom_impl.so
parse-classifier-func-name=NvDsInferParseSolider
classifier-threshold=0.55

yolo_config.txt

[property]
gpu-id=0
net-scale-factor=0.0039215686274509
model-engine-file=yolov8s.onnx_b1_gpu0_fp32.engine
#onnx-file=yolov8n.onnx
batch-size=1
network-mode=0
process-mode=1
num-detected-classes=80
interval=0
#gie-unique-id=2
symmetric-padding=1
cluster-mode=2
parse-bbox-func-name=NvDsInferParseCustomYoloV8
custom-lib-path=parser/libnvdsinfer_custom_impl.so
infer-dims=3;640;640
maintain-aspect-ratio=1
#output-tensor-meta=1

[class-attrs-all]
pre-cluster-threshold=0.75
#topk=20
nms-iou-threshold=0.25

main_config.txt

[application]
enable-perf-measurement=1
perf-measurement-interval-sec=5

[tiled-display]
enable=0

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=3
uri=file:///any-mp4-with-person.mp4
num-sources=1
#drop-frame-interval=2
gpu-id=0
# (0): memtype_device   - Memory type Device
# (1): memtype_pinned   - Memory type Host Pinned
# (2): memtype_unified  - Memory type Unified
cudadec-memtype=0

[sink0]
enable=0
#Type - 1=FakeSink 2=EglSink/nv3dsink (Jetson only) 3=File
type=1
container=1
sync=0
codec=1
enc-type=0
#source-id=1
gpu-id=0
nvbuf-memory-type=0


[sink1]
enable=1
type=3
#1=mp4 2=mkv
container=1
#1=h264 2=h265
codec=1
#encoder type 0=Hardware 1=Software
enc-type=0
#sync=0
#iframeinterval=10
#bitrate=2000000
#H264 Profile - 0=Baseline 2=Main 4=High
#H265 Profile - 0=Main 1=Main10
# set profile only for hw encoder, sw encoder selects profile based on sw-preset
profile=0
output-file=clip2_db.mp4
source-id=0

[osd]
enable=1
gpu-id=0
border-width=1
text-size=10
text-color=1;1;1;1;
text-bg-color=0.3;0.3;0.3;1
font=Serif
show-clock=0
clock-x-offset=800
clock-y-offset=820
clock-text-size=12
clock-color=1;0;0;0
nvbuf-memory-type=0
display-bbox=1
display-text=1

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=0
batch-size=1
## Set muxer output width and height
width=1920
height=1080
##Enable to maintain aspect ratio wrt source, and allow black borders, works
##along with width, height properties
enable-padding=1
nvbuf-memory-type=0
## If set to TRUE, system timestamp will be attached as ntp timestamp
## If set to FALSE, ntp timestamp from rtspsrc, if available, will be attached


[primary-gie]
enable=1
gpu-id=0
batch-size=1
#Required by the app for OSD, not a plugin property
bbox-border-color0=0;0;1;1
bbox-border-color1=0;0;1;1
bbox-border-color2=0;0;1;1
bbox-border-color3=0;0;1;1
interval=0
#Required by the app for SGIE, when used along with config-file property
gie-unique-id=1
nvbuf-memory-type=0
config-file=yolo_config.txt

[secondary-gie1]
enable=1
gpu-id=0
batch-size=1
gie-unique-id=5
input-tensor-meta=0
operate-on-gie-id=1
nvbuf-memory-type=0
config-file=solider_config.txt

[tracker]
enable=1
# For NvDCF and DeepSORT tracker, tracker-width and tracker-height must be a multiple of 32, respectively
tracker-width=640
tracker-height=640
ll-lib-file=/opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
# ll-config-file required to set different tracker types
# ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_IOU.yml
#ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_NvDCF_perf.yml
ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_NvSORT.yml
# ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_NvDCF_accuracy.yml
# ll-config-file=/opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_DeepSORT.yml
#gpu-id=0
display-tracking-id=1

[tests]
file-loop=0