malleshamdasari / fastmot-copy Goto Github PK

/home/clb/github/FastMOT-2.0.0/fastmot/models/osnet_x0_25_msmt17.trt
2021-12-23 10:38:38 [ INFO] Building engine with batch size: 16
2021-12-23 10:38:38 [ INFO] This may take a while...
[TensorRT] ERROR: (Unnamed Layer* 486) [Shuffle]: at most one dimension may be inferred
2021-12-23 10:38:38 [CRITICAL] Failed to parse the ONNX file
2021-12-23 10:38:38 [ ERROR] In node -1 (scaleHelper): UNSUPPORTED_NODE: Assertion failed: dims.nbDims == 4 || dims.nbDims == 5
Traceback (most recent call last):
File "app.py", line 101, in
main()
File "app.py", line 53, in main
mot = fastmot.MOT(config.resize_to, **vars(config.mot_cfg), draw=draw)
File "/home/clb/github/FastMOT-2.0.0/fastmot/mot.py", line 88, in init
self.extractor = FeatureExtractor(**vars(feature_extractor_cfg))
File "/home/clb/github/FastMOT-2.0.0/fastmot/feature_extractor.py", line 28, in init
self.backend = TRTInference(self.model, self.batch_size)
File "/home/clb/github/FastMOT-2.0.0/fastmot/utils/inference.py", line 63, in init
raise RuntimeError('Unable to load the engine file')

jetson agx

import tensorrt as rt
rt.version
'7.2.0.14'

onnx==1.4.1

help thanks