Deepstream-pose-classification-app is not working (deepstream 7.1)

Accelerated Computing Intelligent Video Analytics DeepStream SDK
,
1

• Hardware Platform: Jetson orin nano
• DeepStream Version: 7.1
• JetPack Version: 6.2
• TensorRT Version: v100300
• NVIDIA GPU Driver Version (valid for GPU only): 540.4.0

i am trying to run (deepstream-pose-classification release/tao_ds7.1ga) by following to this topic, but i am facing to this issue:

deepstream_tao_apps/apps/tao_others/deepstream-pose-classification$ ./deepstream-pose-classification-app ../../../configs/app/deepstream_pose_classification_config.yaml
width 1280 hight 720
video file:///home/hbai/Documents/sky-metrics-ai/uploads/Full-second-half.mp4
WARNING: Overriding infer-config batch-size (2) with number of sources (1)
config_file_path:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_preprocess_bodypose_classification.txt
sink_type:2, enc_type:1

(deepstream-pose-classification-app:2991307): GLib-GObject-WARNING **: 13:06:59.317: value "0" of type 'guint' is invalid or out of range for property 'bitrate' of type 'guint'

 *** nv-rtspsink: Launched RTSP Streaming at rtsp://localhost:8554/ds-test ***

Now playing!
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.468999759 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<bodypose-classification-nvinference-engine> NvDsInferContext[UID 4]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 4]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/poseclassificationnet/st-gcn_3dbp_nvidia.onnx_b2_gpu0_fp16.engine
INFO: [FullDims Engine Info]: layers num: 2
0   INPUT  kFLOAT input           3x300x34x1      min: 1x3x300x34x1    opt: 2x3x300x34x1    Max: 2x3x300x34x1    
1   OUTPUT kFLOAT fc_pred         6               min: 0               opt: 0               Max: 0               

0:00:00.469165717 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<bodypose-classification-nvinference-engine> NvDsInferContext[UID 4]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 4]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/poseclassificationnet/st-gcn_3dbp_nvidia.onnx_b2_gpu0_fp16.engine
0:00:00.475117393 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<bodypose-classification-nvinference-engine> [UID 4]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_infer_third_bodypose_classification.txt sucessfully
frameSeqLen:300
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.633745782 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<secondary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 1]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/bodypose3dnet/bodypose3dnet_accuracy.onnx_b8_gpu0_fp16.engine
INFO: [FullDims Engine Info]: layers num: 9
0   INPUT  kFLOAT input0          3x256x192       min: 1x3x256x192     opt: 8x3x256x192     Max: 8x3x256x192     
1   INPUT  kFLOAT k_inv           3x3             min: 1x3x3           opt: 8x3x3           Max: 8x3x3           
2   INPUT  kFLOAT t_form_inv      3x3             min: 1x3x3           opt: 8x3x3           Max: 8x3x3           
3   INPUT  kFLOAT scale_normalized_mean_limb_lengths 36              min: 1x36            opt: 8x36            Max: 8x36            
4   INPUT  kFLOAT mean_limb_lengths 36              min: 1x36            opt: 8x36            Max: 8x36            
5   OUTPUT kFLOAT pose2d          34x3            min: 0               opt: 0               Max: 0               
6   OUTPUT kFLOAT pose2d_org_img  34x3            min: 0               opt: 0               Max: 0               
7   OUTPUT kFLOAT pose25d         34x4            min: 0               opt: 0               Max: 0               
8   OUTPUT kFLOAT pose3d          34x3            min: 0               opt: 0               Max: 0               

0:00:00.633914268 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<secondary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 1]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/bodypose3dnet/bodypose3dnet_accuracy.onnx_b8_gpu0_fp16.engine
0:00:00.638777328 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<secondary-nvinference-engine> [UID 1]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose3d_tao/config_infer_secondary_bodypose3dnet.txt sucessfully
gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
[NvMultiObjectTracker] Loading TRT Engine for tracker ReID...
[NvMultiObjectTracker] Loading Complete!
[NvMultiObjectTracker] Initialized
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.839222941 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 1]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/peoplenet/resnet34_peoplenet_int8.onnx_b2_gpu0_int8.engine
INFO: [FullDims Engine Info]: layers num: 3
0   INPUT  kFLOAT input_1:0       3x544x960       min: 1x3x544x960     opt: 2x3x544x960     Max: 2x3x544x960     
1   OUTPUT kFLOAT output_cov/Sigmoid:0 3x34x60         min: 0               opt: 0               Max: 0               
2   OUTPUT kFLOAT output_bbox/BiasAdd:0 12x34x60        min: 0               opt: 0               Max: 0               

0:00:00.839349186 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 1]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/peoplenet/resnet34_peoplenet_int8.onnx_b2_gpu0_int8.engine
0:00:00.842100968 2991307 0xaaab0ef57430 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<primary-nvinference-engine> [UID 1]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/peoplenet_tao/config_infer_primary_peoplenet.txt sucessfully
Decodebin child added: source
Decodebin child added: decodebin0
Running...
Decodebin child added: qtdemux0
Decodebin child added: multiqueue0
Decodebin child added: h264parse0
Decodebin child added: capsfilter1
Decodebin child added: aacparse0
Decodebin child added: avdec_aac0
Decodebin child added: nvv4l2decoder0
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
In cb_newpad
###Decodebin pick nvidia decoder plugin.
In cb_newpad


**PERF : FPS_0 (14.00)
**PERF : FPS_0 (14.49)
**PERF : FPS_0 (14.33)
**PERF : FPS_0 (13.49)
**PERF : FPS_0 (13.19)
**PERF : FPS_0 (12.66)
**PERF : FPS_0 (12.42)
**PERF : FPS_0 (12.62)
**PERF : FPS_0 (12.44)
**PERF : FPS_0 (12.39)
**PERF : FPS_0 (12.27)
**PERF : FPS_0 (12.41)
**PERF : FPS_0 (12.53)
**PERF : FPS_0 (12.49)
**PERF : FPS_0 (12.59)
**PERF : FPS_0 (12.68)
ResetObjectCtx, object_id:2
**PERF : FPS_0 (12.76)
**PERF : FPS_0 (12.88)
**PERF : FPS_0 (12.89)
**PERF : FPS_0 (13.24)
**PERF : FPS_0 (13.56)
ResetObjectCtx, object_id:1
**PERF : FPS_0 (13.58)
**PERF : FPS_0 (13.73)
**PERF : FPS_0 (13.90)
**PERF : FPS_0 (13.99)
**PERF : FPS_0 (13.99)
**PERF : FPS_0 (13.95)
/dvs/git/dirty/git-master_linux/nvutils/nvbufsurftransform/nvbufsurftransform_copy.cpp:438: => Failed in mem copy

GPUassert: an illegal memory access was encountered /dvs/git/dirty/git-master_linux/deepstream/sdk/src/utils/nvmultiobjecttracker/src/modules/cuDCFv2/cuDCF.cu 944

!![Exception] GPUassert failed
An exception occurred. GPUassert failed
gstnvtracker: Low-level tracker lib returned error 1
ERROR: [TRT]: IExecutionContext::enqueueV3: Error Code 1: Myelin ([cask.cpp:exec:1306] Platform (Cuda) error)
ERROR: Failed to enqueue trt inference batch
ERROR: Infer context enqueue buffer failed, nvinfer error:NVDSINFER_TENSORRT_ERROR
0:00:30.148826694 2991307 0xaaab0e8b5b60 WARN                 nvinfer gstnvinfer.cpp:2115:gst_nvinfer_process_tensor_input:<bodypose-classification-nvinference-engine> error: Failed to queue input batch for inferencing
ERROR from element bodypose-classification-nvinference-engine: Failed to queue input batch for inferencing
ERROR: [TRT]: IExecutionContext::enqueueV3: Error Code 1: Cask (Cask convolution execution)
ERROR: Failed to enqueue trt inference batch
ERROR: Infer context enqueue buffer failed, nvinfer error:NVDSINFER_TENSORRT_ERROR
Error details: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(2115): gst_nvinfer_process_tensor_input (): /GstPipeline:deepstream_pose_classfication_app/GstNvInfer:bodypose-classification-nvinference-engine
0:00:30.149198067 2991307 0xaaab0e8b5de0 WARN                 nvinfer gstnvinfer.cpp:1420:gst_nvinfer_input_queue_loop:<primary-nvinference-engine> error: Failed to queue input batch for inferencing
Returned, stopping playback
[cuOSD Error] at /dvs/git/dirty/git-master_linux/deepstream/sdk/src/utils/nvll_osd/cuda/cuosd_kernel.cu:1072 : Launch kernel (render_elements_kernel) failed, code = 700CUDA Runtime error cudaPeekAtLastError() # an illegal memory access was encountered, code = cudaErrorIllegalAddress [ 700 ] in file /dvs/git/dirty/git-master_linux/deepstream/sdk/src/utils/nvll_osd/cuosd.cpp:968
0:00:30.149575265 2991307 0xaaab0e8b5d80 WARN                 nvinfer gstnvinfer.cpp:2423:gst_nvinfer_output_loop:<primary-nvinference-engine> error: Internal data stream error.
0:00:30.149608355 2991307 0xaaab0e8b5d80 WARN                 nvinfer gstnvinfer.cpp:2423:gst_nvinfer_output_loop:<primary-nvinference-engine> error: streaming stopped, reason error (-5)
ERROR: Failed to make stream wait on event, cuda err_no:700, err_str:cudaErrorIllegalAddress
ERROR: Preprocessor transform input data failed., nvinfer error:NVDSINFER_CUDA_ERROR
0:00:30.150815183 2991307 0xaaab0e8b5800 WARN                 nvinfer gstnvinfer.cpp:1420:gst_nvinfer_input_queue_loop:<secondary-nvinference-engine> error: Failed to queue input batch for inferencing
ERROR: [TRT]: IExecutionContext::enqueueV3: Error Code 1: Cuda Runtime (an illegal memory access was encountered)
ERROR: Failed to enqueue trt inference batch
ERROR: Infer context enqueue buffer failed, nvinfer error:NVDSINFER_TENSORRT_ERROR
0:00:30.151297121 2991307 0xaaab0e8b5b60 WARN                 nvinfer gstnvinfer.cpp:2115:gst_nvinfer_process_tensor_input:<bodypose-classification-nvinference-engine> error: Failed to queue input batch for inferencing
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[ERROR] 2025-09-23 13:07:29 Error destroying cuda device: �����
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[ERROR] 2025-09-23 13:07:29 Error destroying cuda device: ����
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
[WARN ] 2025-09-23 13:07:29 (cudaErrorIllegalAddress)
terminate called after throwing an instance of 'nv::cuda::RuntimeException'
  what():  cudaErrorIllegalAddress: 
Aborted (core dumped)

to fix that i added:

g_object_set(G_OBJECT(nvvideoconvert), "compute-hw", 1, NULL);
g_object_set(G_OBJECT(nvvideoconvert), "nvbuf-memory-type", 3, NULL);

properties to all nvvideoconvert plugin and also runned Makefile to rebuild deepstream_pose_classification_app.cpp, but now i am facing this issue:

deepstream_tao_apps/apps/tao_others/deepstream-pose-classification$ ./deepstream-pose-classification-app ../../../configs/app/deepstream_pose_classification_config.yaml
width 1280 hight 720
video file:///home/hbai/Documents/sky-metrics-ai/uploads/Full-second-half.mp4
WARNING: Overriding infer-config batch-size (2) with number of sources (1)
config_file_path:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_preprocess_bodypose_classification.txt
sink_type:2, enc_type:1

(deepstream-pose-classification-app:2994421): GLib-GObject-WARNING **: 13:18:42.451: value "0" of type 'guint' is invalid or out of range for property 'bitrate' of type 'guint'

 *** nv-rtspsink: Launched RTSP Streaming at rtsp://localhost:8554/ds-test ***

Now playing!
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.462977435 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<bodypose-classification-nvinference-engine> NvDsInferContext[UID 4]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 4]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/poseclassificationnet/st-gcn_3dbp_nvidia.onnx_b2_gpu0_fp16.engine
INFO: [FullDims Engine Info]: layers num: 2
0   INPUT  kFLOAT input           3x300x34x1      min: 1x3x300x34x1    opt: 2x3x300x34x1    Max: 2x3x300x34x1    
1   OUTPUT kFLOAT fc_pred         6               min: 0               opt: 0               Max: 0               

0:00:00.463121152 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<bodypose-classification-nvinference-engine> NvDsInferContext[UID 4]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 4]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/poseclassificationnet/st-gcn_3dbp_nvidia.onnx_b2_gpu0_fp16.engine
0:00:00.469091323 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<bodypose-classification-nvinference-engine> [UID 4]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_infer_third_bodypose_classification.txt sucessfully
frameSeqLen:300
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.622375539 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<secondary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 1]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/bodypose3dnet/bodypose3dnet_accuracy.onnx_b8_gpu0_fp16.engine
INFO: [FullDims Engine Info]: layers num: 9
0   INPUT  kFLOAT input0          3x256x192       min: 1x3x256x192     opt: 8x3x256x192     Max: 8x3x256x192     
1   INPUT  kFLOAT k_inv           3x3             min: 1x3x3           opt: 8x3x3           Max: 8x3x3           
2   INPUT  kFLOAT t_form_inv      3x3             min: 1x3x3           opt: 8x3x3           Max: 8x3x3           
3   INPUT  kFLOAT scale_normalized_mean_limb_lengths 36              min: 1x36            opt: 8x36            Max: 8x36            
4   INPUT  kFLOAT mean_limb_lengths 36              min: 1x36            opt: 8x36            Max: 8x36            
5   OUTPUT kFLOAT pose2d          34x3            min: 0               opt: 0               Max: 0               
6   OUTPUT kFLOAT pose2d_org_img  34x3            min: 0               opt: 0               Max: 0               
7   OUTPUT kFLOAT pose25d         34x4            min: 0               opt: 0               Max: 0               
8   OUTPUT kFLOAT pose3d          34x3            min: 0               opt: 0               Max: 0               

0:00:00.622536953 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<secondary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 1]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/bodypose3dnet/bodypose3dnet_accuracy.onnx_b8_gpu0_fp16.engine
0:00:00.627555473 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<secondary-nvinference-engine> [UID 1]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose3d_tao/config_infer_secondary_bodypose3dnet.txt sucessfully
gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
[NvMultiObjectTracker] Loading TRT Engine for tracker ReID...
[NvMultiObjectTracker] Loading Complete!
[NvMultiObjectTracker] Initialized
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.827421334 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 1]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/peoplenet/resnet34_peoplenet_int8.onnx_b2_gpu0_int8.engine
INFO: [FullDims Engine Info]: layers num: 3
0   INPUT  kFLOAT input_1:0       3x544x960       min: 1x3x544x960     opt: 2x3x544x960     Max: 2x3x544x960     
1   OUTPUT kFLOAT output_cov/Sigmoid:0 3x34x60         min: 0               opt: 0               Max: 0               
2   OUTPUT kFLOAT output_bbox/BiasAdd:0 12x34x60        min: 0               opt: 0               Max: 0               

0:00:00.827535738 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 1]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/peoplenet/resnet34_peoplenet_int8.onnx_b2_gpu0_int8.engine
0:00:00.830107385 2994421 0xaaaaff367430 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<primary-nvinference-engine> [UID 1]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/peoplenet_tao/config_infer_primary_peoplenet.txt sucessfully
Decodebin child added: source
Decodebin child added: decodebin0
Running...
Decodebin child added: qtdemux0
Decodebin child added: multiqueue0
Decodebin child added: h264parse0
Decodebin child added: capsfilter1
Decodebin child added: aacparse0
Decodebin child added: avdec_aac0
Decodebin child added: nvv4l2decoder0
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
In cb_newpad
###Decodebin pick nvidia decoder plugin.
In cb_newpad
/dvs/git/dirty/git-master_linux/nvutils/nvbufsurftransform/nvbufsurftransform.cpp:4543: => Surface type not supported for transformation NVBUF_MEM_CUDA_UNIFIED

ERROR from element qtdemux0: Internal data stream error.
Error details: ../gst/isomp4/qtdemux.c(6760): gst_qtdemux_loop (): /GstPipeline:deepstream_pose_classfication_app/GstBin:source-bin-00/GstURIDecodeBin:uri-decode-bin/GstDecodeBin:decodebin0/GstQTDemux:qtdemux0:
streaming stopped, reason error (-5)
Returned, stopping playback
Deleting pipeline
objects size:20
[NvMultiObjectTracker] De-initialized
2

deepstream_pose_classification_config.yaml


source-list:
   list: file:///Full-second-half.mp4

streammux:
  width: 1280
  height: 720
  batched-push-timeout: 40000

tracker:
  enable: 1
  ll-lib-file: /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
  ll-config-file: /opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_NvDCF_accuracy.yml

primary-gie:
  plugin-type: 0
  config-file-path: ../nvinfer/peoplenet_tao/config_infer_primary_peoplenet.txt
  #config-file-path: ../triton/peoplenet_tao/config_infer_primary_peoplenet.yml
  #config-file-path: ../triton-grpc/peoplenet_tao/config_infer_primary_peoplenet.yml

secondary-gie0:
  plugin-type: 0
  config-file-path: ../nvinfer/bodypose3d_tao/config_infer_secondary_bodypose3dnet.txt
  #config-file-path: ../triton/bodypose3d_tao/config_infer_secondary_bodypose3dnet.yml
  #config-file-path: ../triton-grpc/bodypose3d_tao/config_infer_secondary_bodypose3dnet.yml

secondary-preprocess1:
  config-file-path: ../nvinfer/bodypose_classification_tao/config_preprocess_bodypose_classification.txt

secondary-gie1:
  plugin-type: 0
  config-file-path: ../nvinfer/bodypose_classification_tao/config_infer_third_bodypose_classification.txt
  #config-file-path: ../triton/bodypose_classification_tao/config_infer_third_bodypose_classification.yml
  #config-file-path: ../triton-grpc/bodypose_classification_tao/config_infer_third_bodypose_classification.yml

sink:
  #0 fakesink 
  #1 filesink generate the out.mp4 file in the current directory
  #2 rtspsink publish at rtsp://localhost:8554/ds-test
  #3 displaysink
  sink-type: 2
  #encoder type 0=Hardware 1=Software
  enc-type: 1
4

This is an issue of Jetpack 6.2, Please refer to this FAQ as a workaround.

5

I set compute-hw, copy-hw, nvbuf-memory-type plugins

  g_object_set(G_OBJECT(nvvidconv), "compute-hw", 1, NULL);
  g_object_set(G_OBJECT(nvvidconv), "copy-hw", 2, NULL);
  g_object_set(G_OBJECT(nvvidconv), "nvbuf-memory-type", 3, NULL);

but still same issue.

...
Decodebin child added: multiqueue0
Decodebin child added: h264parse0
Decodebin child added: capsfilter1
Decodebin child added: aacparse0
Decodebin child added: avdec_aac0
Decodebin child added: nvv4l2decoder0
Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NvMMLiteBlockCreate : Block : BlockType = 261 
In cb_newpad
###Decodebin pick nvidia decoder plugin.
In cb_newpad
/dvs/git/dirty/git-master_linux/nvutils/nvbufsurftransform/nvbufsurftransform.cpp:4543: => Surface type not supported for transformation NVBUF_MEM_CUDA_UNIFIED

ERROR from element qtdemux0: Internal data stream error.
Error details: ../gst/isomp4/qtdemux.c(6760): gst_qtdemux_loop (): /GstPipeline:deepstream_pose_classfication_app/GstBin:source-bin-00/GstURIDecodeBin:uri-decode-bin/GstDecodeBin:decodebin0/GstQTDemux:qtdemux0:
streaming stopped, reason error (-5)
Returned, stopping playback
Deleting pipeline
objects size:20
[NvMultiObjectTracker] De-initialized
6

If it can’t be fixed without changing the jetpack version, is deepstream 8.0 compatible with jetpack 6.2?

7

Try setting only the properties listed in the FAQ, as these properties may conflict on Orin.

Currently deepstream-8.0 is only compatible with jetpack-7.0. Currently jetpack-7.0 only supports thor. We will support orin again in subsequent versions.

8

I tried only setting the copy-hw property and running it,

g_object_set(G_OBJECT(ds_source_struct->nvvidconv), "copy-hw", 2, NULL);

but still same issue

nvbufsurftransform_copy.cpp:438: => Failed in mem copy
9

It seems that this workaround does not work on Orin nano. Other users reported that re-flashing JP-6.1 can solve this problem.

10

actually I have my other pipelines written in python, if I install JP-6.1 will my pipelines written in python work?

Suggestion: It would be very convenient if a version of deepstream-pose-classification written in Python was released in subsequent versions.

Thanks

11

Normally, it will work fine.

You can refer to the pyservice-maker demo we provide

12

I think it only supports deepstream version 8.0

image
image1000×180 9.44 KB

13

i am trying to run pyservice-maker pipeline which i copied from deepstream_pose_classification_app.py:

# Copyright and license information
#
# SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: LicenseRef-NvidiaProprietary
#
# NVIDIA CORPORATION, its affiliates and licensors retain all intellectual
# property and proprietary rights in and to this material, related
# documentation and any modifications thereto. Any use, reproduction,
# disclosure or distribution of this material and related documentation
# without an express license agreement from NVIDIA CORPORATION or
# its affiliates is strictly prohibited.
#

from pyservicemaker import Pipeline, Flow, BatchMetadataOperator, Probe, osd
from pyservicemaker.flow import RenderMode
from multiprocessing import Process
import os, sys
from enum import Enum
import math


sys.path.append("../../")
from common.utils import flow_cls_hook, dump_config

# logger = get_logger("deepstream_pose_classification_app")
# logger.setLevel("DEBUG")

class Sink_Type(Enum):
    fakesink = 0
    filesink = 1
    rtspsink = 2
    displaysink = 3
#Padding due to AR SDK model requires bigger bboxes
PAD_DIM=128
_pad_dim=PAD_DIM
MUXER_OUTPUT_WIDTH=1280
MUXER_OUTPUT_HEIGHT=720
_image_width=MUXER_OUTPUT_WIDTH
_image_height=MUXER_OUTPUT_HEIGHT

class ObjectCounterMarker(BatchMetadataOperator):
    def handle_metadata(self, batch_meta):
        # Padding due to AR SDK model requires bigger bounding boxes
        muxer_output_width_pad = _pad_dim * 2 + _image_width
        muxer_output_height_pad = _pad_dim * 2 + _image_height

        for frame_meta in batch_meta.frame_items:
            for obj_meta in frame_meta.object_items:
                sizex = obj_meta.rect_params.width * 0.5
                sizey = obj_meta.rect_params.height * 0.5
                centrx = obj_meta.rect_params.left + sizex
                centry = obj_meta.rect_params.top + sizey
                sizex *= 1.25
                sizey *= 1.25
                if sizex < sizey:
                    sizex = sizey
                else:
                    sizey = sizex
                obj_meta.rect_params.width = round(2.0 * sizex)
                obj_meta.rect_params.height = round(2.0 * sizey)
                obj_meta.rect_params.left = round(centrx - obj_meta.rect_params.width / 2.0)
                obj_meta.rect_params.top = round(centry - obj_meta.rect_params.height / 2.0)

                sizex = obj_meta.rect_params.width * 0.5
                sizey = obj_meta.rect_params.height * 0.5
                centrx = obj_meta.rect_params.left + sizex
                centry = obj_meta.rect_params.top + sizey

                x_scale = 192.0 / sizex
                y_scale = 256.0 / sizey

                if x_scale < y_scale:
                    sizey = 256.0 / x_scale  # Expand height
                else:
                    sizex = 192.0 / y_scale  # Expand width

                obj_meta.rect_params.width = round(2.0 * sizex)
                obj_meta.rect_params.height = round(2.0 * sizey)
                obj_meta.rect_params.left = round(centrx - obj_meta.rect_params.width / 2.0)
                obj_meta.rect_params.top = round(centry - obj_meta.rect_params.height / 2.0)

                if obj_meta.rect_params.left < 0.0:
                    obj_meta.rect_params.left = 0.0
                if obj_meta.rect_params.top < 0.0:
                    obj_meta.rect_params.top = 0.0
                if obj_meta.rect_params.left + obj_meta.rect_params.width > muxer_output_width_pad - 1:
                    obj_meta.rect_params.width = muxer_output_width_pad - 1 - obj_meta.rect_params.left
                if obj_meta.rect_params.top + obj_meta.rect_params.height > muxer_output_height_pad - 1:
                    obj_meta.rect_params.height = muxer_output_height_pad - 1 - obj_meta.rect_params.top


# alias flow_cls_hook as overlay
make_link_element = flow_cls_hook

def sink_out(flow, sink_type, enc_name="test.mp4", enc_type=1, enc_bitrate=2000000):
    print("sink_out", sink_type, enc_name, enc_type, enc_bitrate)
    if sink_type == Sink_Type.fakesink.value:
        flow = flow.render(mode=RenderMode.DISCARD, enable_osd=False, sync=True)
    elif sink_type == Sink_Type.filesink.value:
        flow = flow.encode(dest=enc_name, use_sw_codec=enc_type, bitrate=enc_bitrate)
    elif sink_type == Sink_Type.rtspsink.value:
        flow = flow.make_link_element(["nvrtspoutsinkbin"], [{"enc-type": enc_type}])
    elif sink_type == Sink_Type.displaysink.value:
        flow = flow.render(enable_osd=False)

def deepstream_pose_classification_app(yaml_config_path):
    conf = dump_config(yaml_config_path)
    if conf is None:
        print(f"Failed to load configuration from {yaml_config_path}")
        sys.exit(1)

    pipeline = Pipeline("deepstream_pose_classification_app")
    # add Monkey patch for osd
    Flow.make_link_element = make_link_element
    flow = Flow(pipeline)

    #the pipeline is ......->nvvideoconvert->capsfilter->pgie->tracker->sgie0->nvdspostprocess->
    #->nvdspreprocess->
    flow = flow.batch_capture(
        conf.stream_list, width=1280, height=720
    )

    _pad_dim = PAD_DIM * _image_width / MUXER_OUTPUT_WIDTH
    muxer_output_width_pad = int(_pad_dim * 2 + _image_width)
    muxer_output_height_pad = int(_pad_dim * 2 + _image_height)
    flow = flow.make_link_element(["nvvideoconvert", "capsfilter"],
        [{
        "dest-crop":  f"{_pad_dim}:{_pad_dim}:{_image_width}:{_image_height}",
        "interpolation-method":  1},
         {"caps": f"video/x-raw(memory:NVMM),width={muxer_output_width_pad},height={muxer_output_height_pad}"}
        ]
    )

    flow = flow.infer(conf.pgie_infer_config, with_triton=conf.pgie_infer_type)
    flow = flow.track(ll_config_file=conf.tracker_ll_cfg_file, ll_lib_file=conf.tracker_ll_lib_file)
    flow = flow.attach(what=Probe("counter", ObjectCounterMarker()))
    flow = flow.infer(conf.sgie_0_infer_config, with_triton=conf.sgie_0_infer_type)
    print(conf.postprocess_0_config_file_path, conf.postprocess_0_lib_name)
    flow = flow.make_link_element(["nvdspreprocess", "nvdspostprocess"],
        [{"config-file":  conf.preprocess_1_config_file_path},
         {"postprocesslib-config-file": conf.postprocess_0_config_file_path,
            "postprocesslib-name": conf.postprocess_0_lib_name}]
    )
    flow = flow.infer(conf.sgie_1_infer_config, with_triton=conf.sgie_1_infer_type)

    source_num = len(conf.stream_list)
    tiler_rows = int(math.sqrt(source_num))
    tiler_columns = int(math.ceil(source_num / tiler_rows))
    flow = flow.make_link_element(
        ["nvdslogger", "nvvideoconvert", "capsfilter", "nvmultistreamtiler", "nvvideoconvert", "nvdsosd"],
        [{"fps-measurement-interval-sec":  1},
         { "src-crop":  f"video/x-raw(memory:NVMM),width={muxer_output_width_pad},height={muxer_output_height_pad}"},
         { "caps":  f"video/x-raw(memory:NVMM),width={_image_width},height={_image_height}"},
         { "rows": tiler_rows, "columns": tiler_columns, "width": MUXER_OUTPUT_WIDTH, "height": MUXER_OUTPUT_HEIGHT},
         {},
         {"display-mask": False, "display-bbox": True, "display-text": True, "process-mode": 0},
        ]
    )

    #sink
    sink_out(flow, conf.sinkType, "out.mp4", conf.sinkEncType)

    # Execute the pipeline
    flow()


if __name__ == "__main__":
    # Check input arguments
    if len(sys.argv) != 2:
        print(f"usage: {sys.argv[0]} <yaml> ")
        sys.exit(1)

    # Flow()() is a blocking call due to which the KeyboardInterrupt may not be processed immediately.
    # we use Process from multiprocessing which runs the main function in a different process and processes KeyboardInterrupt immediately.
    process = Process(target=deepstream_pose_classification_app, args=(sys.argv[1],))
    try:
        process.start()
        process.join()
    except KeyboardInterrupt:
        print("\nCtrl+C detected. Terminating process...")
        process.terminate()

but when i run it, i facing this error:

deepstream_tao_apps$ python pysm-apps/deepstream_pose_classification_app.py /deepstream_tao_apps/configs/app/deepstream_pose_classification_config.yaml
Initializing GStreamer Backend...!
Add Element ... batch_capture-mux-1
Add Element ... batch_capture-source-0_0
LINKING: Source: batch_capture-source-0_0 Target: batch_capture-mux-1
Add Element ... flow_cls_hook-nvvideoconvert-0
LINKING: Source: batch_capture-mux-1 Target: flow_cls_hook-nvvideoconvert-0
Add Element ... flow_cls_hook-capsfilter-1
LINKING: Source: flow_cls_hook-nvvideoconvert-0 Target: flow_cls_hook-capsfilter-1
Add Element ... infer-infer-0
Property with-triton is not supported by object infer-infer-0
LINKING: Source: flow_cls_hook-capsfilter-1 Target: infer-infer-0
Add Element ... track-tracker-0
LINKING: Source: infer-infer-0 Target: track-tracker-0
['sgie', '0', 'infer', 'config']
sgie_key: secondary-gie0
['sgie', '0', 'infer', 'type']
sgie_key: secondary-gie0
Add Element ... infer-infer-1
Property with-triton is not supported by object infer-infer-1
LINKING: Source: track-tracker-0 Target: infer-infer-1
['preprocess', '1', 'config', 'file', 'path']
preprocess_key: secondary-preprocess1
['postprocess', '0', 'config', 'file', 'path']
postprocess_key: secondary-postprocess0
Warning: secondary-postprocess0 not found in config, returning None
['postprocess', '0', 'lib', 'name']
postprocess_key: secondary-postprocess0
Warning: secondary-postprocess0 not found in config, returning None
Add Element ... flow_cls_hook-nvdspreprocess-2
LINKING: Source: infer-infer-1 Target: flow_cls_hook-nvdspreprocess-2
Add Element ... flow_cls_hook-nvdspostprocess-3
LINKING: Source: flow_cls_hook-nvdspreprocess-2 Target: flow_cls_hook-nvdspostprocess-3
['sgie', '1', 'infer', 'config']
sgie_key: secondary-gie1
['sgie', '1', 'infer', 'type']
sgie_key: secondary-gie1
Add Element ... infer-infer-2
Property with-triton is not supported by object infer-infer-2
LINKING: Source: flow_cls_hook-nvdspostprocess-3 Target: infer-infer-2
Add Element ... flow_cls_hook-nvdslogger-4
LINKING: Source: infer-infer-2 Target: flow_cls_hook-nvdslogger-4
Add Element ... flow_cls_hook-nvvideoconvert-5
LINKING: Source: flow_cls_hook-nvdslogger-4 Target: flow_cls_hook-nvvideoconvert-5
Add Element ... flow_cls_hook-capsfilter-6
LINKING: Source: flow_cls_hook-nvvideoconvert-5 Target: flow_cls_hook-capsfilter-6
Add Element ... flow_cls_hook-nvmultistreamtiler-7
LINKING: Source: flow_cls_hook-capsfilter-6 Target: flow_cls_hook-nvmultistreamtiler-7
Add Element ... flow_cls_hook-nvvideoconvert-8
LINKING: Source: flow_cls_hook-nvmultistreamtiler-7 Target: flow_cls_hook-nvvideoconvert-8
Add Element ... flow_cls_hook-nvdsosd-9
LINKING: Source: flow_cls_hook-nvvideoconvert-8 Target: flow_cls_hook-nvdsosd-9
sink_out 2 out.mp4 1 2000000
Add Element ... flow_cls_hook-nvrtspoutsinkbin-10
LINKING: Source: flow_cls_hook-nvdsosd-9 Target: flow_cls_hook-nvrtspoutsinkbin-10

(python:3185959): GLib-GObject-CRITICAL **: 15:12:20.942: value "0" of type 'guint' is invalid or out of range for property 'bitrate' of type 'guint'

 *** flow_cls_hook-nvrtspoutsinkbin-10: Launched RTSP Streaming at rtsp://localhost:8554/ds-test ***

Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:00.478178471 3185959 0xffff5a6db9f0 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<infer-infer-2> NvDsInferContext[UID 4]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 4]: deserialized trt engine from :/home/hbai/Documents/deepstream_tao_apps/models/poseclassificationnet/st-gcn_3dbp_nvidia.onnx_b2_gpu0_fp16.engine
INFO: [FullDims Engine Info]: layers num: 2
0   INPUT  kFLOAT input           3x300x34x1      min: 1x3x300x34x1    opt: 2x3x300x34x1    Max: 2x3x300x34x1    
1   OUTPUT kFLOAT fc_pred         6               min: 0               opt: 0               Max: 0               

0:00:00.478291050 3185959 0xffff5a6db9f0 INFO                 nvinfer gstnvinfer.cpp:684:gst_nvinfer_logger:<infer-infer-2> NvDsInferContext[UID 4]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 4]: Use deserialized engine model: /home/hbai/Documents/deepstream_tao_apps/models/poseclassificationnet/st-gcn_3dbp_nvidia.onnx_b2_gpu0_fp16.engine
0:00:00.484229906 3185959 0xffff5a6db9f0 INFO                 nvinfer gstnvinfer_impl.cpp:343:notifyLoadModelStatus:<infer-infer-2> [UID 4]: Load new model:/home/hbai/Documents/deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_infer_third_bodypose_classification.txt sucessfully
FALSE: cannot open shared object file: No such file or directory
Unable to set the pipeline to the playing state.
14

here is deepstream_pose_classification_config.yaml

################################################################################
# SPDX-FileCopyrightText: Copyright (c) 2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: LicenseRef-NvidiaProprietary
#
# NVIDIA CORPORATION, its affiliates and licensors retain all intellectual
# property and proprietary rights in and to this material, related
# documentation and any modifications thereto. Any use, reproduction,
# disclosure or distribution of this material and related documentation
# without an express license agreement from NVIDIA CORPORATION or
# its affiliates is strictly prohibited.
################################################################################

source-list:
   list: file:///Full-second-half-fixed.mp4

streammux:
  width: 1280
  height: 720
  batched-push-timeout: 40000

tracker:
  enable: 1
  ll-lib-file: /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
  ll-config-file: /opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app/config_tracker_NvDCF_accuracy.yml

primary-gie:
  plugin-type: 0
  config-file-path: /deepstream_tao_apps/configs/nvinfer/peoplenet_tao/config_infer_primary_peoplenet.txt
  #config-file-path: ../triton/peoplenet_tao/config_infer_primary_peoplenet.yml
  #config-file-path: ../triton-grpc/peoplenet_tao/config_infer_primary_peoplenet.yml

secondary-gie0:
  plugin-type: 0
  config-file-path: /deepstream_tao_apps/configs/nvinfer/bodypose3d_tao/config_infer_secondary_bodypose3dnet.txt
  #config-file-path: ../triton/bodypose3d_tao/config_infer_secondary_bodypose3dnet.yml
  #config-file-path: ../triton-grpc/bodypose3d_tao/config_infer_secondary_bodypose3dnet.yml

secondary-preprocess1:
  config-file-path: /deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_preprocess_bodypose_classification.txt

secondary-gie1:
  plugin-type: 0
  config-file-path: /deepstream_tao_apps/configs/nvinfer/bodypose_classification_tao/config_infer_third_bodypose_classification.txt
  #config-file-path: ../triton/bodypose_classification_tao/config_infer_third_bodypose_classification.yml
  #config-file-path: ../triton-grpc/bodypose_classification_tao/config_infer_third_bodypose_classification.yml

sink:
  #0 fakesink 
  #1 filesink generate the out.mp4 file in the current directory
  #2 rtspsink publish at rtsp://localhost:8554/ds-test
  #3 displaysink
  sink-type: 2
  #encoder type 0=Hardware 1=Software
  enc-type: 1
15

After a simple modification of pysm_apps deepstream-pose-classification.py using main branch of deepstream_tao_apps, I got the correct result. Please ensure the sample can work fine.

16

I know deepstream-pose-classification.py using main branch of deepstream_tao_apps works properly in deepstream 8.0 version, but as i said before i am using deepstream 7.1 and JP-6.2, so i just copied deepstream-pose-classification.py and common to my local machine and modified common/utils.py to support deepstream 7.1, and run, but i got above error

17

as you can see there is a difference between configs/app/deepstream_pose_classification_config.yaml / master brancha (for deepstream 8.0) and
configs/app/deepstream_pose_classification_config.yaml / releas/tao_ds7.1ga (for deepstream 7.1)

thats why i modified pysm-apps/common/utils.py to support configs/app/deepstream_pose_classification_config.yaml / releas/tao_ds7.1ga (for deepstream 7.1)

18

Yes, I tested deepstream-pose-classification.py on ds-7.1/jp-6.2, Of course, due to environmental differences, some simple modifications are needed.

19

can you please share what kind of modifications did you made and which file did you change

thanks

20 
diff --git a/apps/tao_others/deepstream-pose-classification/postprocesslib_impl/Makefile b/apps/tao_others/deepstream-pose-classification/postprocesslib_impl/Makefile
index 58d8957..f7ad05c 100644
--- a/apps/tao_others/deepstream-pose-classification/postprocesslib_impl/Makefile
+++ b/apps/tao_others/deepstream-pose-classification/postprocesslib_impl/Makefile
@@ -34,7 +34,7 @@ SRCS:= postprocesslib_impl.cpp post_processor.cpp post_processor_bodypose.cpp
 INCS:= $(wildcard *.h)
 LIB:=libpostprocess_impl.so
 
-NVDS_VERSION:=8.0
+NVDS_VERSION:=7.1
 
 CFLAGS+= -fPIC -DDS_VERSION=\"8.0.0\" \
 	 -I /usr/local/cuda-$(CUDA_VER)/include \
diff --git a/pysm-apps/common/utils.py b/pysm-apps/common/utils.py
index 0d79eda..280ddc0 100644
--- a/pysm-apps/common/utils.py
+++ b/pysm-apps/common/utils.py
@@ -13,13 +13,10 @@
 
 import os, yaml
 from pyservicemaker import Flow
-from pyservicemaker.logging import get_logger
 from typing import Dict, Optional, Any
 from cuda import cuda
 import subprocess
 
-logger = get_logger("tao_seg_app")
-logger.setLevel("DEBUG")
 
 
 class Config:
diff --git a/pysm-apps/tao_others/deepstream-pose-classification/deepstream_pose_classification_app.py b/pysm-apps/tao_others/deepstream-pose-classification/deepstream_pose_classification_app.py
index 340c733..5bcc0b3 100644
--- a/pysm-apps/tao_others/deepstream-pose-classification/deepstream_pose_classification_app.py
+++ b/pysm-apps/tao_others/deepstream-pose-classification/deepstream_pose_classification_app.py
@@ -12,7 +12,6 @@
 #
 
 from pyservicemaker import Pipeline, Flow, BatchMetadataOperator, Probe, osd
-from pyservicemaker.logging import get_logger
 from pyservicemaker.flow import RenderMode
 from multiprocessing import Process
 import os, sys
@@ -22,8 +21,6 @@ import math
 sys.path.append("../../")
 from common.utils import flow_cls_hook, dump_config
 
-logger = get_logger("deepstream_pose_classification_app")
-logger.setLevel("DEBUG")
 
 class Sink_Type(Enum):
     fakesink = 0

Just delete some api only supported by ds-8.0