Hai ,
i need to run FaceDetect in jetson xavier NX to detect faces in deepstream testapp1 ,For that i have downloaded the model
The downloaded zip contains model.etlt and int8_calibration.txt files.
I am just confused what all changes i need to make in this configuration file to work.
I had tried multiple times but failed
THESE ARE TH CHANGES I HAVE MADE IN THE CONFIGURATION FILE…WHAT I NEED TO DO WITH THE TEXT FILE ALONG WITH THE MODEL PLEASE HELP
------------------------------------------------------------------------------------------------------------------------------------------- THANKS IN ADVANCE
Thanksss
I have another question. I want to use DeepStream for iris detection, specifically the round part inside the eye. I would like to perform custom training using NVIDIA TAO. Could you please suggest which pretrained model I should use for training?
Or is it possible to train a model in TAO without using pretrained model(From scratch)
hi the code has the referred to facenet engine file which i didn’t find in the respiratory
model-engine-file=…/…/models/faciallandmark/facenet.etlt_b1_gpu0_int8.engine
where can i find this file?
As per FaceDetect TAO page that the model is supported for DS 6.0, can it still work with 6.2?
I’m trying to do something similar to Test-1 sample. so i have only to add the below line as it is and it will be auto generated?
file=…/…/models/faciallandmark/facenet.etlt_b1_gpu0_int8.engine
and if i’m using another TRT file with another name than facenet.etlt i have to modify the line to
file=…/…/FileName.etlt_b1_gpu0_int8.engine
Please refer to the link I attached, the name is related to the configuration of many other parameters. You can comment out this field and wait for the engine file to be generated.
This is the config file
model-file=/opt/nvidia/deepstream/deepstream-6.2/sources/deepstream_tao_apps/models/faciallandmark/facenet.etlt
#proto-file=…/…/…/…/samples/models/Primary_Detector/resnet10.prototxt
model-engine-file=/opt/nvidia/deepstream/deepstream-6.2/sources/deepstream_tao_apps/models/faciallandmark/facenet.etlt_b1_gpu0_int8.engine
labelfile-path=/opt/nvidia/deepstream/deepstream-6.2/samples/models/tao_pretrained_models/deepstream_tao_apps/configs/facial_tao/labels_facenet.txt
int8-calib-file=/opt/nvidia/deepstream/deepstream-6.2/sources/deepstream_tao_apps/models/faciallandmark/int8_calibration.txt
You should make sure there is facenet.etlt file in the /opt/nvidia/deepstream/deepstream-6.2/sources/deepstream_tao_apps/models/faciallandmark/ path. Please read the README of deepstream_tao_apps. We suggest that you do not modify the configuration file first and follow the instructions to run that app.
Hi @axnet , could you open a new topic and describe the detailed information of your environment about your problem? The problem you encountered is not the same as this topic.
