TLT FasterRCNN Confidence Score with tlt-infer

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Description

I am currently working with Transfer Learning Toolkit for FasterRCNN and I am running tlt-infer to run inferencing on my data. I am trying to output the confidence score output. I have the score outputting with the bounding box with the image but I was wondering if I could get it in a text document or with the label dump output.

Environment

TensorRT Version:
GPU Type:
Nvidia Driver Version:
CUDA Version:
CUDNN Version:
Operating System + Version:
Python Version (if applicable):
TensorFlow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if container which image + tag):

Relevant Files

Please attach or include links to any models, data, files, or scripts necessary to reproduce your issue. (Github repo, Google Drive, Dropbox, etc.)

Steps To Reproduce

Please include:

  • Exact steps/commands to build your repro
  • Exact steps/commands to run your repro
  • Full traceback of errors encountered
1 Like
2

Hi, Request you to share the model, script, profiler and performance output so that we can help you better.

Alternatively, you can try running your model with trtexec command
https://github.com/NVIDIA/TensorRT/tree/master/samples/opensource/trtexec
or view these tips for optimizing performance
https://docs.nvidia.com/deeplearning/tensorrt/best-practices/index.html

Thanks!

3

Attached below are the inference configs on the experiment spec sheet

inference_config {
image_dir: ‘/ze/data/Experiments/DAB/D_1/images’
model: ‘C:/Users/Quinn/zeroEyes/src/ZE-TLT/T_123/models/model.epoch12.tlt’
batch_size: 1
detection_image_output_dir: ‘/ze/data/Experiments/DAB/D_1/inference/images’
labels_dump_dir: ‘/ze/data/Experiments/DAB/D_1/inference/labels’
rpn_pre_nms_top_N: 6000
rpn_nms_max_boxes: 300
rpn_nms_overlap_threshold: 0.700000
bbox_visualize_threshold: 0.600000
classifier_nms_max_boxes: 300
classifier_nms_overlap_threshold: 0.300000
bbox_caption_on: True
}

here is the command line arguement

#tlt-infer faster_rcnn -e /ze/data/experiments/TLT/T_2/specs/network_spec

here are sample label dumps from the inference

1 0 0 0 861.56 9.86 903.43 75.19 0 0 0 0 0 0 0

The confidence can be scene in the inference image dump along with the bounding box and class, but the value is not located in the corresponding label dump. The confidence of the specific label above was .7403

4

Hi @quinn,
I may suggest you to raise the concern in TLT forum to get better assistance.
Thanks!