[Issue report] input-tensor-meta or input-tensor-from-meta

• Hardware Platform (GPU)
• DeepStream Version 6.0
• JetPack Version (valid for Jetson only)
• TensorRT Version 8.0.1
• NVIDIA GPU Driver Version 11.4
• Issue Type( questions)
**• How to reproduce the issue ? **
set input-tensor-meta=1 in nvinfer config.txt
then see
“Unknown or legacy key specified ‘input-tensor-meta’ for group [property]”.
After trace the source code and find
#define CONFIG_GROUP_INFER_INPUT_FROM_META “input-tensor-from-meta”
So I changed input-tensor-from-meta=1 in nvinfer config.txt
Suggest to update NvInfer document
Or change naming constant CONFIG_GROUP_INFER_INPUT_FROM_META

I’ve checked the source code in the DeepStream SDK 6.0 GA package. The definition is #define CONFIG_GROUP_GIE_INPUT_TENSOR_META “input-tensor-meta” in /opt/nvidia/deepstream/deepstream-6.0/sources/apps/apps-common/src/deepstream_config_file_parser.c

Dear Fiona,
Thanks for your reply.
My DeepStream docker image is "FROM nvcr.io/nvidia/deepstream:6.0-devel"
In gstnvinfer_property_parser.cpp
I found gst_nvinfer_parse_props() only parse 
CONFIG_GROUP_INFER_INPUT_FROM_META 
but not CONFIG_GROUP_GIE_INPUT_TENSOR_META.
I think that's why I met 
“Unknown or legacy key specified ‘input-tensor-meta’ for group [property]”.
Could you confirm it? Thanks.

Yes. I got your idea now. We will check and update the document.