Deepstream failed when using TLT model as pgie

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) = dGPU
• DeepStream Version = 5.0
• JetPack Version (valid for Jetson only)
• TensorRT Version = 7
**• NVIDIA GPU Driver Version (valid for GPU only) = 450.36.06 **

I used peoplenet model (https://ngc.nvidia.com/catalog/models/nvidia:tlt_peoplenet) as pgie for deepstream_infer_tensor_meta_test.cpp

The config file for pgie is below

[property]

gpu-id=0

net-scale-factor=0.0039215697906911373

tlt-model-key=tlt_encode

tlt-encoded-model=/opt/nvidia/deepstream/deepstream-5.0/samples/models/tlt_pretrained_models/peoplenet/resnet34_peoplenet_pruned.etlt

labelfile-path=/opt/nvidia/deepstream/deepstream-5.0/samples/configs/tlt_pretrained_models/labels_peoplenet.txt

# model-engine-file=/opt/nvidia/deepstream/deepstream-5.0/samples/models/tlt_pretrained_models/peoplenet/resnet34_peoplenet_pruned.etlt_b1_gpu0_fp32.engine

input-dims=3;544;960;0

uff-input-blob-name=input_1

batch-size=1

process-mode=1

model-color-format=0

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

network-mode=0

num-detected-classes=3

cluster-mode=1

interval=0

gie-unique-id=1

output-blob-names=output_bbox/BiasAdd;output_cov/Sigmoid

[class-attrs-all]

pre-cluster-threshold=0.4

## Set eps=0.7 and minBoxes for cluster-mode=1(DBSCAN)

eps=0.7

minBoxes=1

I got the following error message when I run it.

0:00:11.330941648  6060 0x557d3c97faa0 INFO                 nvinfer gstnvinfer_impl.cpp:311:notifyLoadModelStatus:<primary-nvinference-engine> [UID 1]: Load new model:dstensor_pgie_config_peoplenet.txt sucessfully
Running...

** (deepstream-infer-tensor-meta-app:6060): ERROR **: 09:46:48.498: failed to check pgie network info

Trace/breakpoint trap

Please help,
Thanks!

The same config file for pgie (peoplenet TLT model) is working with deepstream test2 sample app.

@vorapolbig

We are looking into this issue and setting up similar environments to reproduce this fault.
I may ask you for extra information if needed later.

Thanks you.

Hi,

any update on the issue?

@ersheng

have you tried to reproduce the fault?

@vorapolbig

Sorry for replying so late.
In the C++ code, the macro definitions

#define PGIE_NET_WIDTH 640
#define PGIE_NET_HEIGHT 368

should agree with input-dims=3;544;960;0 defined in dstensor_pgie_config_peoplenet.txt