Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

Upgrade tensorflow to 2.1.0 #28

Merged
merged 4 commits into from Jan 17, 2020

Conversation

dyeo
Copy link

@dyeo dyeo commented Jan 17, 2020

馃帀 Issues resolved:

馃И How to Test

  1. Update your CUDA install to v10.1 and install the latest cuDNN for that version
  2. Restart your computer
  3. Update your installation and dependencies by running pip install .
  4. Run the game as normal

Caveats

  • Upgrading to tensorflow 2.1.0 results in this issue on windows 10 devices, even with the correct CUDA version installed. The suggested fix is to downgrade to tensorflow 14.0.0, which obviously doesn't benefit us.
    • This seems to be a non-issue.

Copy link
Contributor

@dustinlacewell dustinlacewell left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't really have enough ML background to review the semantics of these changes, but the Python looks fine. Let's wait for a review by @WAUthethird to get it merged.

Copy link
Contributor

@WAUthethird WAUthethird left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Well done! Seems this upgrade addresses all the issues and conversion factors needed, however I'm unsure about requiring users to install CUDA. Seems that Microsoft Visual C++ Redistributable for Visual Studio 2015 needs to be installed to solve the error, as per this link: tensorflow/tensorflow#22794 (comment)

Will test and report back.

@WAUthethird
Copy link
Contributor

I've tested on TF 2.1 (most recent version) and TF 2.0 and I don't receive CUDA errors, rather the expected warnings in the initialization process. I have the Visual C++ Rd for VS 2015 installed. CUDA is not required.

You can safely update the package versions. I've confirmed that tensorflow 2.1.0 and tensorflow-estimator 2.0.1 work with this PR with modification to the version checks. tensorboard can most likely be updated without issue.

@dyeo
Copy link
Author

dyeo commented Jan 17, 2020

Updated to 2.1.0 馃殌

@dyeo dyeo changed the title Upgrade tensorflow to 2.0.0 Upgrade tensorflow to 2.1.0 Jan 17, 2020
@WAUthethird WAUthethird merged commit 0cbd79d into storybro:develop Jan 17, 2020
@WAUthethird WAUthethird mentioned this pull request Jan 17, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants