ValueError: Unsupported data type 14 in tensor

Gaurav_Reddy · December 9, 2021, 9:34pm

System information

Have I written custom code : interpreter = tf.lite.Interpreter(model_path="models/saved_model.tflite")
OS Platform and Distribution : RaspberryPi 3B+
TensorFlow Git version : v2.4.0-0-g582c8d2
TensorFlow version : 2.4.0
Installation Source : https://github.com/bitsy-ai/tensorflow-arm-bin/releases/download/v2.4.0/tensorflow-2.4.0-cp37-none-linux_armv7l.whl
Python version: Python 3.7.3
More Info about the tflite Model : Custom trained EfficientDet-B0 on 40 classes

ERROR :

Traceback (most recent call last):
  File "detect.py", line 8, in <module>
    interpreter = tf.lite.Interpreter(model_path="models/saved_model.tflite")
  File "/home/pi/.local/lib/python3.7/site-packages/tensorflow_core/lite/python/interpreter.py", line 209, in __init__
    model_path, self._custom_op_registerers))
ValueError: Unsupported data type 14 in tensor
Unsupported data type 14 in tensor
Unsupported data type 14 in tensor
Unsupported data type 14 in tensor
Unsupported data type 14 in tensor
Unsupported data type 14 in tensor

Note : If you are suggesting me to upgrade or downgrade the packages, please provide necessary code with it.

Thanks !!

chunduriv · July 27, 2023, 10:34am

@Gaurav_Reddy,

You can try to clone Tensorflow Repo and checkout 2.13 version which is the latest stable version and run the below command

PYTHON=python3 tensorflow/lite/tools/pip_package/build_pip_package_with_cmake.sh native

Please follow these instructions for the installation.

Thank you!