GstNvInfer:nvinfer_bin_nvinfer: Internal data stream error

Accelerated Computing Intelligent Video Analytics DeepStream SDK
,
1

• Jetson Orin Nano
• DS 7.0.0
• JP 6.0.0 v2

Running d0_avlp_544_960_f32.engine from github.com/NVIDIA-AI-IOT/deepstream_tao_apps.
Config file, except I am not converting the model to int8.

What could be the troubleshooting steps? Thanks.

2

Hi,

Internal data stream error is the default GStreamer error when any element in the pipeline returns something other than GST_FLOW_OK. To get more information, you should enable GStreamer debug logging: GST_DEBUG=4 gst-launch-1.0 ...

We also have a dedicated wiki page for GStreamer debugging if you need more details:

Feel free to share your logs with GST_DEBUG enabled, and I’ll check if it’s an error I recognize.

4

are you testing efficientdet model with nvinfer? what do you mean about “except I am not converting the model to int8”? could you share simplified code and cfg diff? could you share the detailed reproducing steps and logs? Thanks!

5

Thanks for your responses guys!

gxf_server_logs_2609.txt.gz (122.1 KB)

gzip -d gxf_server_logs_2609.txt.gz

@fanzh

ds_gc_graph.tar.gz (2.3 KB)

I have a DeepStream graph for the Graph Composer, here are the contents of the archive:

  • graph_config.yaml - defines the graph in Graph Composer;
  • graph_params.yaml - overrides the params from graph_config.yaml;
  • trafficcamnet_config.txt - config for the 1st detector;
  • lpd_edet_de.txt - config for the 2nd (cascade/downstream) detector.

Is there anything else I could provide?

6

from the log, the fatal error is “0:03:18.158370353 3836 0xffff68003de0 ERROR egladaption gstegladaptation.c:689:gst_egl_adaptation_choose_config:<video_renderer/video_renderer-sink> Could not find matching framebuffer config”
it is related to egl display. please refer to topic1 and topic2.

7

Solved, thanks!

8

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