An error occurred during deployment

Diluk · March 9, 2021, 3:02am

Hi，I got the engine file using tlt-converter on the Jetson platform, but the following error occurred when I ran it:

My configuration file is as follows：

Thanks.

Morganh · March 9, 2021, 6:00am

Did you train the frcnn model with TLT 2.0 or TLT 3.0?

Diluk · March 9, 2021, 6:39am

I am using TLT2.0

Morganh · March 9, 2021, 6:42am

For TLT 2.0, please follow https://docs.nvidia.com/metropolis/TLT/archive/tlt-20/tlt-user-guide/text/deploying_to_deepstream.html#integrating-a-fasterrcnn-model
and the config file in https://github.com/NVIDIA-AI-IOT/deepstream_tlt_apps/tree/release/tlt2.0.1/nvdsinfer_customparser_frcnn_tlt

Diluk · March 10, 2021, 9:45am

I carefully read the README in deepstream_tlt_apps, compiled and replaced the oss file in jetson, when I executed
‘./ds-tlt -c …/configs/frcnn_tlt/pgie_frcnn_tlt_config.txt -i ./sample_720p.h264’,
it appeared The following error,

any idea? thanks

Morganh · March 10, 2021, 10:37am

Please check if you have the write access to the folder.

Diluk · March 11, 2021, 3:21am

I successfully ran the sample, but when I replaced it with the etlt file I trained, the following error occurred：

Below is my pgie_frcnn_tlt_config.txt

Morganh · March 11, 2021, 6:22am

Please double check the config file according to the error prompt.