What 's the version of the keras + tensorflow

[CristXu]

Describe the bug

After built and install the onnx-tf by pip install -e .
The onnx-tf is installed correctly, but after run the onn-tf,
it will said: no module keras.src.engin
My package version:
keras 2.12.0
tensorboard 2.16.2
tensorboard-data-server 0.7.2
tensorflow 2.16.1
tensorflow-addons 0.22.0
tensorflow-intel 2.16.1
tensorflow-io-gcs-filesystem 0.31.0
tensorflow-probability 0.24.0
termcolor 2.4.0
tf_keras 2.16.0

To Reproduce

Please give us instructions to reproduce your problem.

A self-contained piece of code that can demonstrate the problem is required.

Please do not expect us to have PyTorch, Caffe2 installed.

If a model exported from PyTorch and Caffe2 is having trouble in ONNX-TF, use the next section to attach the model.

ONNX model file

If applicable, attach the onnx model file in question using Gist, DropBox or Google Drive.

Python, ONNX, ONNX-TF, Tensorflow version

This section can be obtained by running get_version.py from util folder.

• Python version:
• ONNX version:
• ONNX-TF version:
• Tensorflow version:

Additional context

Add any other context about the problem here.

[umihesh]

I was able to convert a model from ONNX format to tf, and tf to tf-lite with the following dependency versions:

onnx==1.10.2
onnx_tf==1.10.0
tensorflow==2.8.0
tensorflow-addons==0.17.1
tensorflow-probability==0.16.0
numpy==1.20.0
protobuf==3.19.6 #(For tf -> tf-lite conversion)

Installing the TensorFlow dependencies will also get you the correct tf-keras (2.15.0) and keras (2.8.0) versions.

In theory, python versions 3.8.x - 3.9.x would work, but practically I tested it with 3.9.19.
Note also that you need GPU to run the conversion. I have no idea why, but it only worked this way for me.

[PoornavG]

I was able to convert a model from ONNX format to tf, and tf to tf-lite with the following dependency versions:

onnx==1.10.2
onnx_tf==1.10.0
tensorflow==2.8.0
tensorflow-addons==0.17.1
tensorflow-probability==0.16.0
numpy==1.20.0
protobuf==3.19.6 #(For tf -> tf-lite conversion)

Installing the TensorFlow dependencies will also get you the correct tf-keras (2.15.0) and keras (2.8.0) versions.

What version of python did you use ?

[umihesh]

I was able to convert a model from ONNX format to tf, and tf to tf-lite with the following dependency versions:

onnx==1.10.2
onnx_tf==1.10.0
tensorflow==2.8.0
tensorflow-addons==0.17.1
tensorflow-probability==0.16.0
numpy==1.20.0
protobuf==3.19.6 #(For tf -> tf-lite conversion)

Installing the TensorFlow dependencies will also get you the correct tf-keras (2.15.0) and keras (2.8.0) versions.

What version of python did you use ?

In theory, 3.8.x - 3.9.x would work. I tested it with 3.9.19.
Note also that you need GPU to run the conversion. I have no idea why, but it only worked this way for me.