Repository navigation
chore(deps): update nvcr.io/nvidia/cuda docker tag to v13 - #14699
renovate-bot wants to merge 1 commit into
Conversation
trusted-contributions-gcf
Bot
added
kokoro:force-run
product-auto-label
Bot
added
the
samples
kokoro-team
removed
the
kokoro:force-run
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request updates the base Docker images across multiple TensorFlow GPU examples to use CUDA 13.4.3 instead of CUDA 12.5.1. However, this upgrade is incompatible with TensorFlow 2.21.0, which is built against CUDA 12.x, and will break GPU acceleration at runtime. It is recommended to revert these changes to maintain a compatible CUDA 12.x base image.
| # Supported NVIDIA images: https://gitlab.com/nvidia/container-images/cuda/blob/master/doc/supported-tags.md | ||
| # TensorFlow/CUDA compatibility: https://www.tensorflow.org/install/source#gpu | ||
| FROM nvcr.io/nvidia/cuda:12.5.1-cudnn-runtime-ubuntu22.04 | ||
| FROM nvcr.io/nvidia/cuda:13.4.3-cudnn-runtime-ubuntu22.04 |
There was a problem hiding this comment.
Upgrading the base image to CUDA 13.4.3 is incompatible with TensorFlow 2.21.0 (specified in requirements.txt). TensorFlow 2.21.0 is built against CUDA 12.x and will fail to find the expected CUDA 12 shared libraries (such as libcudart.so.12) at runtime, breaking GPU acceleration. Please revert this change to keep a compatible CUDA 12.x base image.
FROM nvcr.io/nvidia/cuda:12.5.1-cudnn-runtime-ubuntu22.04
| # Supported NVIDIA images: https://gitlab.com/nvidia/container-images/cuda/blob/master/doc/supported-tags.md | ||
| # TensorFlow/CUDA compatibility: https://www.tensorflow.org/install/source#gpu | ||
| FROM nvcr.io/nvidia/cuda:12.5.1-cudnn-runtime-ubuntu22.04 | ||
| FROM nvcr.io/nvidia/cuda:13.4.3-cudnn-runtime-ubuntu22.04 |
There was a problem hiding this comment.
Upgrading the base image to CUDA 13.4.3 is incompatible with TensorFlow 2.21.0 (specified in requirements.txt). TensorFlow 2.21.0 is built against CUDA 12.x and will fail to find the expected CUDA 12 shared libraries (such as libcudart.so.12) at runtime, breaking GPU acceleration. Please revert this change to keep a compatible CUDA 12.x base image.
FROM nvcr.io/nvidia/cuda:12.5.1-cudnn-runtime-ubuntu22.04
| # Supported NVIDIA images: https://gitlab.com/nvidia/container-images/cuda/blob/master/doc/supported-tags.md | ||
| # TensorFlow/CUDA compatibility: https://www.tensorflow.org/install/source#gpu | ||
| FROM nvcr.io/nvidia/cuda:12.5.1-cudnn-runtime-ubuntu22.04 | ||
| FROM nvcr.io/nvidia/cuda:13.4.3-cudnn-runtime-ubuntu22.04 |
There was a problem hiding this comment.
Upgrading the base image to CUDA 13.4.3 is incompatible with TensorFlow 2.21.0 (specified in requirements.txt). TensorFlow 2.21.0 is built against CUDA 12.x and will fail to find the expected CUDA 12 shared libraries (such as libcudart.so.12) at runtime, breaking GPU acceleration. Please revert this change to keep a compatible CUDA 12.x base image.
FROM nvcr.io/nvidia/cuda:12.5.1-cudnn-runtime-ubuntu22.04
This PR contains the following updates:
12.5.1-cudnn-runtime-ubuntu22.04→13.4.3-cudnn-runtime-ubuntu22.04Configuration
📅 Schedule: (UTC)
🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.
♻ Rebasing: Never, or you tick the rebase/retry checkbox.
🔕 Ignore: Close this PR and you won't be reminded about this update again.
This PR was generated by Mend Renovate. View the repository job log.