Deep Learning OSS Nvidia Driver AMI GPU PyTorch 2.9 (Amazon Linux 2023) 20251124
For help getting started, see Getting started with DLAMI.
Notices
Starting with PyTorch 2.9, CUDA and related libraries are now located within the Pytorch virtual environment found in
/opt/pytorch/cudainstead of being installed at the system level. This change reduces the image size while maintaining full CUDA functionality. Please ensure that the PyTorch virtual environment is activated to utilize CUDA usingsource /opt/pytorch/bin/activate
Major Package Versions
supported_ec2_instances
|
G4dn, G5, G6, Gr6, G6e, P4, P4de, P5, P5e, P5en, P6-B200
|
operating_system
|
Amazon Linux 2023.9.20251117
|
compute_architecture
|
aarch64
|
kernel_version
|
6.1.158-178.288.amzn2023.aarch64
|
framework_version
|
2.9
|
python_location
|
/opt/pytorch/bin/python
|
nvidia_driver
|
580.95.05
|
gdr_copy
|
2.5.1
|
nvidia_container_toolkit_version
|
1.18.0
|
efa_version
|
1.44.0
|
ofi_nccl_version
|
1.17.1
|
ebs_volume_type
|
gp3
|
ssm_agent_version
|
3.3.3050.0
|
Package Changes from Previous Release
Note
All DLAMI are scanned for vulnerabilities following Amazon Best Practices. While we continuously upgrade to the latest package versions, some dependencies may require specific package downgrades to maintain compatibility. These adjustments do not impact our security posture. For information on Amazon security measures, please refer to our Amazon Security Documentation
Added Packages
No packages were added in this release.
Updated Packages
No packages were updated in this release.
Removed Packages
No packages were removed in this release.