search

ceccocats / tkDNN

Deep neural network library and toolkit to do high performace inference on NVIDIA jetson platforms
GNU General Public License v2.0
718 stars 208 forks source link

CUDNN failure: CUDNN_STATUS_NOT_INITIALIZED #282

Closed ckurtz22 closed 2 years ago

ckurtz22 commented 2 years ago 
tkDNN/build$ ./test_yolo4
Not supported field: batch=1
Not supported field: subdivisions=1
Not supported field: momentum=0.949
Not supported field: decay=0.0005
Not supported field: angle=0
Not supported field: saturation = 1.5
Not supported field: exposure = 1.5
Not supported field: hue=.1
Not supported field: learning_rate=0.00261
Not supported field: burn_in=1000
Not supported field: max_batches = 500500
Not supported field: policy=steps
Not supported field: steps=400000,450000
Not supported field: scales=.1,.1
Not supported field: mosaic=1
New NETWORK (tkDNN v0.5, CUDNN v8)
CUDNN failure: CUDNN_STATUS_NOT_INITIALIZED
/gfr_ws/src/see_camera_processing/libs/tkDNN/src/Network.cpp:56
Aborting...

running on Jetson AGX Xavier in Docker image. Unsure what the issue is, it works fine outside the docker. Spent a few days trying various different methods of running Nv stuff in Docker to no avail.

ckurtz22 commented 2 years ago

So it turns out this issue that has been driving me crazy for 2 weeks is because I was linking my user into the Docker container and not running it as root. ugh.