Environment Details:
• Hardware Platform (Jetson / GPU): dGPU (x86_64)
• DeepStream Version: 7.1
• JetPack Version (valid for Jetson only): N/A
• TensorRT Version: 10.3.0.26
• NVIDIA GPU Driver Version: 560.35.05
• Issue Type: Question
Hello, I am working with DeepStream 7.1 on an x86_64 platform and Graph Composer 4.1.0. I have implemented a simple pipeline, which runs as expected, but I’ve encountered a roadblock when it comes to handling inference metadata.
Issue Details:
- I want to save the inference results (e.g., object classifications, bounding boxes) to a file in JSON or text format.
- So far, I haven’t found a way to extract or log this metadata from the pipeline in Graph Composer.
Questions:
- Is there a built-in node or mechanism in Graph Composer to extract and save inference metadata?
- How can I configure the pipeline to output these results to a file (e.g., JSON or txt)?
- If direct logging isn’t possible, are there any alternative approaches, such as adding a custom node or modifying configurations?
I would greatly appreciate any guidance, examples, or pointers to relevant documentation that could help resolve this issue.
Thank you for your assistance!