Help with running Deepstream (with yoloV3)

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hello! I am very new to Nvidia Deepstream and have followed along the QuickStart Guide step-by-step but could not get it to work.
Jetpack Version: 4.4
Deepstream Verson: 5.1
(I think Jetpack 4.5 now works with Deepstream but I installed it about a month ago when 4.4 was required for Deepstream 5.1)

First of all, my goal is to use it with YoloV3 on a depth-sensing camera, but for now I am testing on a logitech webcam. I was able to run AlexeyAB’s yolov3 by itself with no problem.

For deepstream, I tried multiple config files (such as ones in /samples/configs/deepstream-app, and /sources/objectDetector_Yolo).
Currently, I keep getting the error:

deepstream-app -c config_infer_primary_yoloV3.txt
** ERROR: main:655: Failed to set pipeline to PAUSED
Quitting
ERROR from src_bin_muxer: Output width not set
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvmultistream/gstnvstreammux.c(2512): gst_nvstreammux_change_state (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstNvStreamMux:src_bin_muxer
App run failed

I get this same error for every file, and when I use “vim (config file)”, it looks a lot shorter than ones I saw in example videos. This is what my “config_infer_primary_yoloV3.txt” looks like:

[property]
gpu-id=0
net-scale-factor=0.0039215697906911373
#0=RGB, 1=BGR
model-color-format=0
custom-network-config=yolov3.cfg
model-file=yolov3.weights
#model-engine-file=yolov3_b1_gpu0_int8.engine
labelfile-path=labels.txt
int8-calib-file=yolov3-calibration.table.trt7.0

0=FP32, 1=INT8, 2=FP16 mode

network-mode=1
num-detected-classes=80
gie-unique-id=1
network-type=0
is-classifier=0

0=Group Rectangles, 1=DBSCAN, 2=NMS, 3= DBSCAN+NMS Hybrid, 4 = None(No clustering)

cluster-mode=2
maintain-aspect-ratio=1
parse-bbox-func-name=NvDsInferParseCustomYoloV3
custom-lib-path=nvdsinfer_custom_impl_Yolo/libnvdsinfer_custom_impl_Yolo.so
engine-create-func-name=NvDsInferYoloCudaEngineGet
#scaling-filter=0
#scaling-compute-hw=0

[class-attrs-all]
nms-iou-threshold=0.3
threshold=0.7

In other examples I saw of people opening it, it had many more options such as ‘width, height, etc’ which I guess I am supposed to specify.

Lastly, some places tell me to edit these files, but when I try, it says it’s read-only, and cannot even use “chmod 775” or “chmod 777”.

Am I just not understanding how to properly run the sample files/models/configs in Deepstream?

Thank you for any help, if I missed any vital information please ask for it.

3

Could you upgrade to JP4.5.1 for DS5.1?
Or you can use DS5.0.1 on JP4.4.

image
image1469×491 30.5 KB

4

Thank you for the reply. I tried installing Deepstream 5.0.0 (instead of 5.0.1 since I am following steps from another project that used the same setup, and double checking with the archived guide for 5.0 with dependencies), and when trying to run any config file, I get this:

** ERROR: <create_multi_source_bin:1320>: Failed to create element ‘src_bin_muxer’
** ERROR: <create_multi_source_bin:1411>: create_multi_source_bin failed
** ERROR: <create_pipeline:1294>: create_pipeline failed
** ERROR: main:636: Failed to create pipeline
Quitting
App run failed

I read somewhere it might be a gstreamer problem so I tried using “export GST_PLUGIN_PATH=”/usr/lib/aarch64-linux-gnu/gstreamer-1.0/", which didn’t seem to do anything.

Any solutions I should try? Can’t seem to find much in my case. Much appreciated, thank you!

5

I believe this is your setup issue.
Please check if you install a correct and complete Jetpack 4.4. And, if you will use DS5.0, you can install it via SDKManager.

For refer, you can also refer to Quickstart Guide — DeepStream 6.1.1 Release documentation to install DS5.0 package manually on JP4.4 system.

6

Hmm. Well I am quite sure I installed JP4.4 correctly, I double checked with installation videos and most other programs work fine. I also already have many many files on the SD card and do not wish to wipe it with a reinstallation of Jetpack.
I’m using Jetson Xavier NX and installed DS5.0 with the tar file manually.

Are there any other solutions? Thank you very much.

7

Ok, got your question.
You can’t use config_infer_primary_yoloV3.txt, it’s a sub-config file for GIE only.
You need to use deepstream_app_config_yoloV3.txt instead, and change its source according to your input source.

For logitech webcam, is it a usb camera? If it’s, you should change its source to be like below

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI
type=1
camera-width=640
camera-height=480
camera-fps-n=30
camera-fps-d=1
camera-v4l2-dev-node=0
9

Thank you for the reply.
So I am using an old USB Logitech webcam, I tried using the same file and changing the [source0] to have exactly that, however, i get this error:

deepstream-app -c deepstream_app_config_yoloV3.txt
Unknown or legacy key specified ‘is-classifier’ for group [property]
Warn: ‘threshold’ parameter has been deprecated. Use ‘pre-cluster-threshold’ instead.

Using winsys: x11
gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream-5.0/lib/libnvds_mot_klt.so
gstnvtracker: Optional NvMOT_RemoveStreams not implemented
gstnvtracker: Batch processing is OFF
gstnvtracker: Past frame output is OFF
ERROR: Could not open lib: /opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/nvdsinfer_custom_impl_Yolo/libnvdsinfer_custom_impl_Yolo.so, error string: /opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/nvdsinfer_custom_impl_Yolo/libnvdsinfer_custom_impl_Yolo.so: cannot open shared object file: No such file or directory
0:00:00.351125300 13426 0x7f44002350 ERROR nvinfer gstnvinfer.cpp:613:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Error in NvDsInferContextImpl::initialize() <nvdsinfer_context_impl.cpp:1139> [UID = 1]: Could not open custom lib: (null)
0:00:00.351205813 13426 0x7f44002350 WARN nvinfer gstnvinfer.cpp:809:gst_nvinfer_start:<primary_gie> error: Failed to create NvDsInferContext instance
0:00:00.351230901 13426 0x7f44002350 WARN nvinfer gstnvinfer.cpp:809:gst_nvinfer_start:<primary_gie> error: Config file path: /opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/config_infer_primary_yoloV3.txt, NvDsInfer Error: NVDSINFER_CUSTOM_LIB_FAILED
** ERROR: main:655: Failed to set pipeline to PAUSED
Quitting
ERROR from primary_gie: Failed to create NvDsInferContext instance
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(809): gst_nvinfer_start (): /GstPipeline:pipeline/GstBin:primary_gie_bin/GstNvInfer:primary_gie:
Config file path: /opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo/config_infer_primary_yoloV3.txt, NvDsInfer Error: NVDSINFER_CUSTOM_LIB_FAILED
App run failed

10

Please follow the README to get the default YoloV3 sample running firstly