tensorflow: RuntimeError: Inputs and outputs not all float|uint8|int16 types.Node number 2 (ADD) failed to invoke.

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Ubuntu 18.04
• TensorFlow installed from (source or binary): source
• TensorFlow version (or github SHA if from source): 1.15

Command used to run the converter or code if you’re using the Python API If possible, please share a link to Colab/Jupyter/any notebook.

import numpy as np
import tensorflow as tf
from PIL import Image

interpreter = tf.lite.Interpreter(model_path="out.tflite")
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
print(input_details)
[{'name': 'image', 'index': 21904, 'shape': array([  3, 270, 480], dtype=int32), 'dtype': <class 'numpy.float32'>, 'quantization': (0.0, 0)}]
print(output_details)
[{'name': 'action', 'index': 7204, 'shape': array([], dtype=int32), 'dtype': <class 'numpy.float32'>, 'quantization': (0.0, 0)}]
input_shape = input_details[0]['shape']
input_data = np.array(np.random.random_sample(input_shape), dtype=np.float32)
interpreter.set_tensor(input_details[0]['index'], input_data)
interpreter.invoke()
output_data = interpreter.get_tensor(output_details[0]['index'])
print(output_data)

The output from the converter invocation

RuntimeError: Inputs and outputs not all float|uint8|int16 types.Node number 2 (ADD) failed to invoke.

Also, please include a link to the saved model or GraphDef

https://we.tl/t-lWH3XmYihS <-- .pb
https://we.tl/t-Bkid4ThzN1 <-- .tflite

Failure details The conversion is successful and I can run inference on the .pb, but I get the following error when I run inference on the tflite file.

RuntimeError: Inputs and outputs not all float|uint8|int16 types.Node number 2 (ADD) failed to invoke.

Any other info / logs

Model trained on pytorch mobilenetV2 arch, converted to onnx, and converted to tensorflow.