Yolov4_tiny_usa_deployable for LPD don't detect anything

Accelerated Computing Intelligent Video Analytics DeepStream SDK
, ,
1

• GPU
*• DeepStream 7.0
• CUDA 12.6
• detection bug
• Try a deepstream pipeline with LPDnet model pruned_v2.1

Using yolov4_tiny_usa_deployable for pruned_v2.1 (LPDNet | NVIDIA NGC) on my deepstream pipeline, it detected absolutely nothing, 0 licence plate.

Here my pipeline

pipeline:
  - v4l2src:
      device: /dev/video0
  - capsfilter:
      caps: "image/jpeg, width=1920, height=1080, framerate=30/1"
  - jpegdec: {}
  - videoconvert: {}
  - nvvideoconvert: {}
  - capsfilter:
      caps: "video/x-raw(memory:NVMM), format=RGBA, width=1920, height=1080"
  - mux.sink_0:
       nvstreammux:
          name: mux
          batch-size: 1
          width: 1920
          height: 1080
          batched-push-timeout: 4000000
          live-source: 1
          num-surfaces-per-frame: 1
          sync-inputs: 0
          max-latency: 0
  - nvinfer:
      config-file-path: ../infer_cfg/config_infer_secondary_lpdnet_YOLO.txt
  - nvtracker:
      tracker-width: 640
      tracker-height: 384
      gpu-id: 0
      ll-lib-file: /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
      ll-config-file: ../infer_cfg/config_tracker_NvDCF_perf.yml
  - nvdsanalytics:
      name: "analytics"
      config-file: ../infer_cfg/analytics.txt
  - nvvideoconvert: {}
  - nvdsosd:
      name: onscreendisplay
  - fpsdisplaysink:
      name: fps-display
      video-sink: nveglglessink
      text-overlay: false
      sync: false

Here my config file for the model

[property]
gpu-id=0
net-scale-factor=1.0
offsets=103.939;116.779;123.68
model-color-format=1
labelfile-path=../../models/LP/LPD/usa_lpd_label.txt
tlt-encoded-model=../../models/LP/LPD/yolov4_tiny_usa_deployable.etlt
int8-calib-file=../../models/LP/LPD/yolov4_tiny_usa_cal.bin
tlt-model-key=nvidia_tlt
infer-dims=3;480;640  # Corrected from `uff-input-dims`
batch-size=16
network-mode=1  # INT8 mode (ensure your hardware supports it)
num-detected-classes=1
process-mode=2  # Secondary inference
interval=0
gie-unique-id=2
network-type=0  # Detector type
operate-on-gie-id=1
operate-on-class-ids=0
cluster-mode=3  # NMS clustering mode
output-blob-names=BatchedNMS
parse-bbox-func-name=NvDsInferParseCustomBatchedNMSTLT
custom-lib-path=../../deepstream_tao_apps/post_processor/libnvds_infercustomparser_tao.so
#enable-dla=1  # Uncomment if using Deep Learning Accelerator (DLA)

[class-attrs-all]
pre-cluster-threshold=0.3
roi-top-offset=0
roi-bottom-offset=0
detected-min-w=0
detected-min-h=0
detected-max-w=0
detected-max-h=0

here my output/warnings before launch :

gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
[NvMultiObjectTracker] Initialized
0:00:02.101643023 419733 0x57888284f920 INFO                 nvinfer gstnvinfer.cpp:682:gst_nvinfer_logger:<nvinfer0> NvDsInferContext[UID 2]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:2109> [UID = 2]: Trying to create engine from model files
WARNING: [TRT]: onnx2trt_utils.cpp:374: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
WARNING: [TRT]: onnx2trt_utils.cpp:400: One or more weights outside the range of INT32 was clamped
WARNING: [TRT]: builtin_op_importers.cpp:5221: Attribute caffeSemantics not found in plugin node! Ensure that the plugin creator has a default value defined or the engine may fail to build.
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 125) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 207) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 208) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 210) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 217) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 218) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 224) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 225) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 289) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 290) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 293) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 294) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 298) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 299) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 555) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 556) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 559) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor (Unnamed Layer* 560) [Shuffle]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
WARNING: [TRT]: Missing scale and zero-point for tensor BatchedNMS, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
3

If you are using CUDA 12.6, please upgrade your DeepStream version to 7.1.

About the LPD model, you can try to use our latest version: 2.3.1.