YoloV3 without Batched NMS post processing code

[abhimanyu.kspillai]

I am using TAO Toolkit version 5.3.0 and has trained a YoloV3 model, which on conversion adds the BatchedNMS layer to it, which was not supported on my embedded device.

So I removed the Batched NMS layer as mentioned in this discussion.

Errors while reading ONNX file produced by TAO 5

Please use below way to trim the onnx. For example, to trim https://forums.developer.nvidia.com/uploads/short-url/tc4tmv15H16NiFPOqOa1QNswBnM.onnx import onnx_graphsurgeon as gs import numpy as np import onnx model = onnx.load("cardbox_detection_with_yolov4_tiny.onnx") graph = gs.import_onnx(model) tensors = graph.tensors() graph.inputs = [tensors["Input"].to_variable(dtype=np.float32, shape=( "N", 3,512,768))] graph.outputs = [tensors["box"].to_variable(dtype=np.float32), tensors["cls"].to…

Requesting you help me with the post processing/inference code for the same withou BatchedNMS in it.

Thanks & Regards,
Abhimanyu K S

[Morganh]

@abhimanyu.kspillai

post processing/inference code for the same withou BatchedNMS in it

The official code of the BatchedNMS plugin is in TensorRT/plugin/batchedNMSPlugin at 23.08 · NVIDIA/TensorRT · GitHub.
If the plugin exists, the boxes and scores are the two inputs of the batchedNMSPlugin.
If you already remove the plugin, you can do the post-processing against the boxes and scores directly.

[yingliu]

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks