Cannot use channel_offsets in preprocess in streamer/config/sgie_config.txt

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
• DeepStream Version
• JetPack Version (valid for Jetson only)
• 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)
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

Deepstream 6.2
GPU

4

which sample are you testing? what is the whole media pipeline?
which plugin are you using to do preprocess?

5

plugin: nvinferserver

it throw an error when setting channel_ offsets in preprocess part od the config file:
preprocess {
network_format: IMAGE_FORMAT_BGR
tensor_order: TENSOR_ORDER_LINEAR
maintain_aspect_ratio: 1
frame_scaling_hw: FRAME_SCALING_HW_DEFAULT
frame_scaling_filter: 1
symmetric_padding: 1
normalize {
scale_factor: 1.0
channel_offsets: [77.5, 21.2, 11.8] <----ERROR
}
}

(when I add the line “channel_offsets: [77.5, 21.2, 11.8]” it throw the error:

INFO: infer_grpc_backend.cpp:169 TritonGrpcBackend id:2 initialized for model: resnet
ERROR: infer_cuda_context.cpp:238 failed to set preprocess scalor and offsets.
0:00:00.154625384 11463 0x55bbd1c93380 ERROR nvinferserver gstnvinferserver.cpp:407:gst_nvinfer_server_logger: nvinferserver[UID 2]: Error in initialize() <infer_base_context.cpp:89> [UID = 2]: Infer Context failed to create preprocessors., nvinfer error:NVDSINFER_CONFIG_FAILED
0:00:00.154639188 11463 0x55bbd1c93380 WARN nvinferserver gstnvinferserver_impl.cpp:588:start: error: Failed to initialize InferTrtIsContext
0:00:00.154642314 11463 0x55bbd1c93380 WARN nvinferserver gstnvinferserver_impl.cpp:588:start: error: Config file path: config/sgie_resnet50_config.txt
0:00:00.154810976 11463 0x55bbd1c93380 WARN nvinferserver gstnvinferserver.cpp:509:gst_nvinfer_server_start: error: gstnvinferserver_impl start failed

6

this error is from NetworkPreprocessor::setScaleOffsets, nvinferserver plugin is opensource, please find it in \opt\nvidia\deepstream\deepstream-6.2\sources\libs\nvdsinferserver\infer_cuda_context.cpp.

  1. is the model 's input channel number three?
  2. can you use deepstream sample objectDetector_FasterRCNN to reproduce this issue? noticing there are similar settings in the configuration file \opt\nvidia\deepstream\deepstream-6.2\sources\objectDetector_FasterRCNN\config_triton_inferserver_primary_fasterRCNN_custom.txt.
7

Thanks fanzh,

It’s working now.
I add to change tensor_order from TENSOR_ORDER_LINEAR to TENSOR_ORDER_NHWC which is the same order as the model Inputs.

1 Like
8

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