I used Transfer Learning Toolkit to train resnet18detectnetv2 for custom dataset. I then used the tlt-converter to convert the .etlt model into an .engine file. I am able to deploy both the .etlt and .engine file in DeepStream and it works.
But now I need to deploy the model in python and I am not able to find how to load the .engine model in python.
Hi sharanssundar,
For " deploy the model in python ", it is not the deploy way in current version of TLT.
See https://devtalk.nvidia.com/default/topic/1065558/transfer-learning-toolkit/trt-engine-deployment/ for reference.
For “how to load the .engine model in python”, if you already generate a trt engine from etlt model, then your question becomes a TensorRT topic.
See doc: https://docs.nvidia.com/deeplearning/sdk/tensorrt-sample-support-guide/index.html
There are many python examples, especially under /usr/src/tensorrt/samples/python/
Hi @sharanssundar,
I have some question about train the resnet18detectnetv2 for custom my own dataset.
1- You converted your dataset to kitti dataset format? If so, you resize your images and boxes for fixed size like kitti dataset size?
2- for converting the dataset to tfrecord, tlt-tfrecord-converter expected to have 16 fields in labels text files, you filled other fields any value like zeros?
3- In the detectnet_v2_train_resnet18_kitti.txt, there are all config about classes of kitti, like this:
target_class_config {
key: "pedestrian"
value {
clustering_config {
coverage_threshold: 0.00749999983236
dbscan_eps: 0.230000004172
dbscan_min_samples: 0.0500000007451
minimum_bounding_box_height: 20
}
}
}
target_classes {
name: "pedestrian"
class_weight: 4.0
coverage_foreground_weight: 0.0500000007451
objectives {
name: "cov"
initial_weight: 1.0
weight_target: 1.0
}
objectives {
name: "bbox"
initial_weight: 10.0
weight_target: 10.0
}
}
There are many such config about claases of kitti, I need to modify all of this to my custom dataset classes?
4- Could you run trt engine like .engine file with python code without deep stream sdk? if so, please share your code if possible.
5- How to you deploy both the .etlt and .engine file in DeepStream? on jetson nano or PC? using deep stream docker image or tar file or deb file?