Is Joint model (OF + RGB) for ActionRecognitionNet supported in Deepstream?

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) T4
• DeepStream Version 6.4
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs)
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)

When I try to run a joint Action Recognition Net modeled that I trained using Tao I get an error. There seems to be discrepancies between TAO and Deepstream docs regarding whether it’s supported or not. If Joint models are supported, how do I deploy them? Below is my Deepstream config file.

# deepstream action recognition config settings.
# run:
# $ deepstream-3d-action-recognition -c deepstream_action_recognition_config.txt

[action-recognition]

# stream/file source list
uri-list=file:///app/videos/clip.mp4;

# eglglessink settings
display-sync=1
# 0=eglgles display; 1=fakesink
fakesink=0

# <preprocess-config> is the config file path for nvdspreprocess plugin
# <infer-config> is the config file path for nvinfer plugin

# Enable 3D preprocess and inference
preprocess-config=config_preprocess_3d_custom.txt
infer-config=config_infer_primary_3d_action.txt

# Uncomment to enable 2D preprocess and inference
#preprocess-config=config_preprocess_2d_custom.txt
#infer-config=config_infer_primary_2d_action.txt

# nvstreammux settings
muxer-height=720
muxer-width=1280

# nvstreammux batched push timeout in usec
muxer-batch-timeout=40000


# nvmultistreamtiler settings
tiler-height=720
tiler-width=1280

# Log debug level. 0: disabled. 1: debug. 2: verbose.
debug=1

# Enable fps print on screen. 0: disable. 1: enable
enable-fps=1

• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

When I run it I get this error:

root@instance:/app# deepstream-3d-action-recognition -c deepstream_action_recognition_config_3d.txt
num-sources = 1
Now playing: file:///app/videos/clip.mp4,
0:00:07.944425681   132 0x55fc0bc2aca0 INFO                 nvinfer gstnvinfer.cpp:682:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2092> [UID = 1]: deserialized trt engine from :/app/models/joint_model_epoch019.engine
WARNING: [TRT]: The getMaxBatchSize() function should not be used with an engine built from a network created with NetworkDefinitionCreationFlag::kEXPLICIT_BATCH flag. This function will always return 1.
INFO: ../nvdsinfer/nvdsinfer_model_builder.cpp:610 [Implicit Engine Info]: layers num: 3
0   INPUT  kFLOAT input_rgb       96x224x224      
1   INPUT  kFLOAT input_of        64x224x224      
2   OUTPUT kFLOAT fc_pred         5               

0:00:08.068822876   132 0x55fc0bc2aca0 INFO                 nvinfer gstnvinfer.cpp:682:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2195> [UID = 1]: Use deserialized engine model: /app/models/joint_model_epoch019.engine
0:00:08.072453936   132 0x55fc0bc2aca0 WARN                 nvinfer gstnvinfer.cpp:679:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Warning from NvDsInferContextImpl::initNonImageInputLayers() <nvdsinfer_context_impl.cpp:1611> [UID = 1]: More than one input layers but custom initialization function not implemented
0:00:08.072477897   132 0x55fc0bc2aca0 ERROR                nvinfer gstnvinfer.cpp:676:gst_nvinfer_logger:<primary-nvinference-engine> NvDsInferContext[UID 1]: Error in NvDsInferContextImpl::initialize() <nvdsinfer_context_impl.cpp:1385> [UID = 1]: Failed to initialize non-image input layers
0:00:08.079001984   132 0x55fc0bc2aca0 WARN                 nvinfer gstnvinfer.cpp:898:gst_nvinfer_start:<primary-nvinference-engine> error: Failed to create NvDsInferContext instance
0:00:08.079021877   132 0x55fc0bc2aca0 WARN                 nvinfer gstnvinfer.cpp:898:gst_nvinfer_start:<primary-nvinference-engine> error: Config file path: config_infer_primary_3d_action.txt, NvDsInfer Error: NVDSINFER_CUSTOM_LIB_FAILED
Running...
ERROR from element primary-nvinference-engine: Failed to create NvDsInferContext instance
Error details: gstnvinfer.cpp(898): gst_nvinfer_start (): /GstPipeline:preprocess-test-pipeline/GstNvInfer:primary-nvinference-engine:
Config file path: config_infer_primary_3d_action.txt, NvDsInfer Error: NVDSINFER_CUSTOM_LIB_FAILED
Returned, stopping playback
Deleting pipeline

deepstream-3d-action-recognition sample only support version 1.0 of Action Recognition Net | NVIDIA NGC. The version 2.0 model needs optical flow input, the case is not supported by DeepStream now. Please consult in TAO forum for how to train the version 1.0 model.

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.