DNN library is not found

trying to run object detection inference using tensorflow 2.12 on jetson orin nano 8gb , cuda 11.4 cudnn 8.6 jetpack 5.1.2 , got the following error:

Traceback (most recent call last):
  File "utils/tf2_model/object_detection_cam.py", line 84, in <module>
    detections = detect_fn(input_tensor)
  File "/usr/local/lib/python3.8/dist-packages/tensorflow/python/saved_model/load.py", line 740, in _call_attribute
    return instance.__call__(*args, **kwargs)
  File "/usr/local/lib/python3.8/dist-packages/tensorflow/python/util/traceback_utils.py", line 153, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/usr/local/lib/python3.8/dist-packages/tensorflow/python/eager/execute.py", line 52, in quick_execute
    tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
tensorflow.python.framework.errors_impl.UnimplementedError: Graph execution error:

2 root error(s) found.
  (0) UNIMPLEMENTED:  DNN library is not found.
	 [[{{node ssd_mobile_net_v2keras_feature_extractor/functional_1/Conv1/Conv2D}}]]
  (1) UNIMPLEMENTED:  DNN library is not found.
	 [[{{node ssd_mobile_net_v2keras_feature_extractor/functional_1/Conv1/Conv2D}}]]
0 successful operations.
0 derived errors ignored. [Op:__inference_restored_function_body_52267]

trying to load model : SSD MobileNet v2 320x320 from : models/research/object_detection/g3doc/tf2_detection_zoo.md at master · tensorflow/models · GitHub

more info :

>>> tf.__version__
>>> print("Num GPUs Available: ", len(tf.config.list_physical_devices('GPU')))
Num GPUs Available:  1
>>> tf.test.is_gpu_available()
WARNING:tensorflow:From <stdin>:1: is_gpu_available (from tensorflow.python.framework.test_util) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.config.list_physical_devices('GPU')` instead.
>>> tf.test.is_built_with_cuda()

any idea how to solve this issue ?
Thanks in advance


Have you tried to run the same app on other platforms that have GPU, ex desktop environment?
If not, please give it a try. The experiment helps to check if this issue is Jetson-specific.


@AastaLLL yes it worked on desktop environment ubunutu 20.04 tensorflow 2.12 , cuda 11.8, cudnn9.0

FYI, I tried cuda 11.8 and cudnn9.0 on the jetson still not working

FYI Inference worked on CPU

Can you please attach the relevant code that generates the exception for me to try and replicate?

Hi @guyda
i followed this tutorial :


Would you mind sharing how you set up the environment as well?
We try to reproduce this internally but get stuck by TFOD → tensorflow_text → bazel → jdk → java.lang.NoClassDefFoundError: Could not initialize class com.google.devtools.build.lib.unsafe.StringUnsafe error.