• Hardware Platform (Jetson / GPU): jetson
• DeepStream Version : 5.1
• JetPack Version (valid for Jetson only) : 4.5
• TensorRT Version : 7.1.x
I fine-tuned usa_lpd model for custom dataset, and I want to run that model with DLA on jetson xavier nx , but I don’t know How I can generate bin file for the trained model?
Is it possible to use usa_lpd_dla.bin file for my trained model? Is it correct way?
I converted the trained model to engine file with below command on GTX 1080 ti, and this generated two files, one is engine file and other is bin file.
But when I use that bin file, That’s not work for generating engine with jetson when I want to run DLA, but for jetson’s GPU is ok.
And when I want to use the usa_lpd_dla.bin, the model is generated on DLA.
tlt detectnet_v2 export \ -m $USER_EXPERIMENT_DIR/experiment_dir_retrain/weights/resnet18_detector_pruned.tlt \ -o $USER_EXPERIMENT_DIR/experiment_dir_final/resnet18_detector.etlt \ -k $KEY \ --cal_data_file $USER_EXPERIMENT_DIR/experiment_dir_final/calibration.tensor \ --data_type int8 \ --batches 10 \ --batch_size 4 \ --max_batch_size 4\ --engine_file $USER_EXPERIMENT_DIR/experiment_dir_final/resnet18_detector.trt.int8 \ --cal_cache_file $USER_EXPERIMENT_DIR/experiment_dir_final/calibration.bin \ --verbose