NVIDIA Developer Forums

Custom TAO unet model classifying only two classes on Deepstream!

Accelerated Computing Intelligent Video Analytics TAO Toolkit

Morganh May 12, 2022, 2:58am 35

Please modify threshold to

float threshold = 0.3;

The result is as below.

MAJOR ACCURACY LOSS when EXPORTING tao unet model after retraining pruned model

Topic		Replies	Views	Activity
DeepStream compatibility issues with UNET output layer change TAO Toolkit	27	1654	November 25, 2022
TAO trained 3-class Unet model only outputs 2 class output in Deepstream DeepStream SDK gstreamer , tao , deepstream , computer-vision	7	825	October 9, 2024
Custom UNet segmentation model uses only 2 colors for output map DeepStream SDK tao , deepstream	23	1149	September 23, 2022
Issues with tao classifier_tf2 in deepstream (Accuracy drops) TAO Toolkit deepstream	21	79	September 6, 2024
Custom new model segmentation in DeepStream DeepStream SDK tensorflow	41	1452	February 3, 2023
Deepstream infrence gives no detection TAO Toolkit	28	1947	December 9, 2021
Fine-tuned TAO ClassificationTF2 Accuracy Drop after Compiling to TensorRT TAO Toolkit	34	875	August 6, 2024
Issue with image classification tutorial and testing with deepstream-app TAO Toolkit tensorrt , jetson-inference	34	5866	October 12, 2021
TAO 5.0 Classification (PyTorch) deploy error TAO Toolkit	49	1528	September 11, 2023
Integrating Tao Models (detectnet_v2) into Deepstream SDK TAO Toolkit tao , deepstream , jetson-nano	11	1005	March 24, 2023