Tensorflow 1 is unsupported in Colab

Course Q&A Generative Adversarial Networks (GANS) Build Basic Generative Adversarial Networks
1

I wanted to run (Colab) Pre-trained model exploration.ipynb, but I faced this problem.
Tensorflow 1 is unsupported in Colab.
Your notebook should be updated to use Tensorflow 2.
—> 41 See the guide at Migrate to TensorFlow 2  |  TensorFlow Core“”"
42 ))
43

ValueError: Tensorflow 1 is unsupported in Colab
Would you please help me?

2

Hi @Reza_shafiloo,
See this post for a work-around suggestion until the developers can get this fixed:

1 Like
3

Hi Wendy,
I looked up the workaround and used !pip install tensorflow-gpu==1.15.2. It still doesn’t work… any update on this issue?

4

Hi @sooolee,
I just tried with the work-around and it is still working for me. Is yours failing when you run the second cell? I do see an error about pip’s dependency resolver, but it doesn’t seem to cause any issues when running the rest of the cells in the colab.

This is a good reminder, too. I’ll check with the development team to see how the official fix is coming along.

5

Tried it again, and it worked. Thanks!

1 Like
6

Hi, I am facing this error: UnknownError: 2 root error(s) found.
(0) Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
[[node Gs_1/_Run/Gs/G_synthesis/4x4/Conv/Conv2D (defined at /usr/local/lib/python3.7/dist-packages/tensorflow_core/python/framework/ops.py:1748) ]]
[[Gs_1/_Run/saturate_cast/_2843]]
(1) Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
[[node Gs_1/_Run/Gs/G_synthesis/4x4/Conv/Conv2D (defined at /usr/local/lib/python3.7/dist-packages/tensorflow_core/python/framework/ops.py:1748) ]]
0 successful operations.
0 derived errors ignored.

7

@Sai_Sandeep_Kantared, are you seeing this error after you’ve replaced the

%tensorflow_version 1.x

line with:
!pip install tensorflow-gpu==1.15.2

I’m not able to reproduce the issue. Can you bring up the colab again, make the change above, and then try choosing “Run all” from the Runtime menu to check if you’re still seeing the issue?