Which version of Keras, Tensorflow and Pytorch are compatible?

[asr-lord]

I'm using Google Colab to run this code over CPU and GPU but it doesn't work when I'm trying to run it.

My libraries version:

Keras==2.4.3
tensorflow==2.5.0
torch==1.9.0+cu102

Error message:

python speakerDiarization.py

2021-06-25 12:23:18.495455: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudart.so.11.0
Traceback (most recent call last):
  File "speakerDiarization.py", line 11, in <module>
    import model as spkModel
  File "ghostvlad/model.py", line 28, in <module>
    class VladPooling(keras.engine.Layer):
AttributeError: module 'keras.engine' has no attribute 'Layer'

I tried with older versions of Tensorflow and Keras:

tensorflow==2.1
Keras==2.3.1

but I still getting errors:

Using TensorFlow backend.
2021-06-25 12:29:00.291956: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer.so.6'; dlerror: libnvinfer.so.6: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/lib64-nvidia
2021-06-25 12:29:00.292135: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer_plugin.so.6'; dlerror: libnvinfer_plugin.so.6: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/lib64-nvidia
2021-06-25 12:29:00.292155: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:30] Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
Traceback (most recent call last):
  File "speakerDiarization.py", line 11, in <module>
    import model as spkModel
  File "ghostvlad/model.py", line 3, in <module>
    import keras
  File "/usr/local/lib/python3.7/dist-packages/keras/__init__.py", line 3, in <module>
    from . import utils
  File "/usr/local/lib/python3.7/dist-packages/keras/utils/__init__.py", line 26, in <module>
    from .vis_utils import model_to_dot
  File "/usr/local/lib/python3.7/dist-packages/keras/utils/vis_utils.py", line 7, in <module>
    from ..models import Model
  File "/usr/local/lib/python3.7/dist-packages/keras/models.py", line 10, in <module>
    from .engine.input_layer import Input
  File "/usr/local/lib/python3.7/dist-packages/keras/engine/__init__.py", line 3, in <module>
    from .input_layer import Input
  File "/usr/local/lib/python3.7/dist-packages/keras/engine/input_layer.py", line 7, in <module>
    from .base_layer import Layer
  File "/usr/local/lib/python3.7/dist-packages/keras/engine/base_layer.py", line 12, in <module>
    from .. import initializers
  File "/usr/local/lib/python3.7/dist-packages/keras/initializers/__init__.py", line 124, in <module>
    populate_deserializable_objects()
  File "/usr/local/lib/python3.7/dist-packages/keras/initializers/__init__.py", line 49, in populate_deserializable_objects
    LOCAL.GENERATED_WITH_V2 = tf.__internal__.tf2.enabled()
AttributeError: module 'tensorflow_core.compat.v2' has no attribute '__internal__'

[taylorlu]

tensorflow 1.8-1.15 and pytorch 1.3.0 may work.

[Grolaf]

Hey there !

I installed tensorflow 1.15 and pytorch 1.3.0 but the keras version required by the code is missig, since the last keras version require tensorflow 2.2.0....

Which version of keras do we need ?

Thanks for all !

[taylorlu]

You can take a look at this page https://docs.floydhub.com/guides/environments

[asr-lord]

I have tried it but with no luck.

[taylorlu]

Tensorflow 1.8.0 Keras 2.2.4 pytorch 1.3.0

[tanujdhiman]

Tensorflow 1.8.0 Keras 2.2.4 pytorch 1.3.0

[ghost]

Is this working with python 3.8.x? I'm using python 3.8 and couldn't find tensorflow 1.x version. Any updates would be helpful for Ubuntu 20.04, Python3.8 and tensorflow 2.x