• Hardware Platform ( GPU)
• DeepStream Version 6.3
• TensorRT Version 8.6.1.6
• NVIDIA GPU Driver Version (valid for GPU only) 12.2
I have a classification model for Gender and Age, but during the fine-tuning process. After training with my own data, I’m encountering unsatisfactory results. I’ve been using this model for classification: GitHub - GilLevi/AgeGenderDeepLearning.
Could you please check if anyone has a good model for fine-tuning in a classification pipeline for Gender and age adaptable in DeepStream pipeline ?
Is there any TAO model can be leveraged for this requirement?
There is no similar TAO model for the gender and age classification.
DeepStream can support Caffe Model and Caffe Prototxt, ONNX and TensorRT model. DeepStream is for inferencing only. There is no special limitation of which model you want to use from DeepStream point.
You may consider to train your model with some 3rd party dataset to improve the classification result. insightface/recognition/datasets at master · deepinsight/insightface (github.com)
Thank you. I use these models in Zoo, and when I employ the models’ inferences for classification, I perform some preprocessing. However, in DeepStream, I only utilize a file configuration. Where can I find this preprocessing step in deepstream ?
There is preprocessing in gst-nvinfer plugin. Gst-nvinfer — DeepStream 6.3 Release documentation. The gst-nvinfer plugin is open source, the source code is in /opt/nvidia/deepstream/deepstream/sources/gst-plugins/gst-nvinfer and /opt/nvidia/deepstream/deepstream/sources/libs/nvdsinfer. The diagram here DeepStream SDK FAQ - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums can help you to read the code.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.