Runtime Version List

Cloud Machine Learning Engine uses images, based on runtime version designation, to configure cloud resources to service your training and prediction requests. This page lists the runtime versions and their constituent packages. You can learn more about runtime versions and how to work with them in the versioning overview.

Supported Cloud ML runtime versions

The following versions are supported to run training jobs and serve both online and batch predictions:

Version Package Released On Last Updated Notes
1.6 TensorFlow 1.6
The following PyPI packages:
xgboost 0.7.post3
numpy 1.13.3
pandas 0.19.2
scipy 0.19.1
scikit-learn 0.19.1
sympy 1.0
statsmodels 0.8.0
oauth2client 2.2.0
httplib2 0.9.2
python-dateutil 2.6.0
argparse 1.4.0
six 1.10.0
PyYAML 3.12
wrapt 1.10.10
crcmod 1.7
google-api-python-client 1.6.2
python-json-logger 0.1.7
gcloud 0.18.3
subprocess32 3.2.7 (removed when running in Python 3 training)
wheel 0.30.0a0
WebOb 1.7.2
Paste 2.0.3
tornado 4.5.1
grpcio 1.8.6
requests 2.13.0
webapp2 3.0.0b1
google-cloud-logging 1.0.0
The following Ubuntu packages:
curl
libcurl3-dev
wget
zip
unzip
git
vim
build-essential
ca-certificates
ca-certificates-java
pkg-config
rsync
libatlas-base-dev
liblapack-dev
libpng-dev
libjpeg-dev
gfortran
python3.5
python3-dev
python3-pip
python3-setuptools

python2.7
python-dev
python-setuptools
gdb
openjdk-8-jdk
openjdk-8-jre-headless
g++
zlib1g-dev
libio-all-perl
module-init-tools
libsnappy-dev
libyaml-0-2
python-opencv
March 20, 2018 April 5, 2018

Runtime version 1.6 supports TensorFlow 1.6.0 for CPU and GPU.

Python 3.5 is available for training and online prediction with runtime version 1.6. The Ubuntu packages for Python 3 (indicated in bold) are installed when running Python 3.

1.5 TensorFlow 1.5
The following PyPI packages:
xgboost 0.7.post3
numpy 1.13.3
pandas 0.19.2
scipy 0.19.1
scikit-learn 0.19.1
sympy 1.0
statsmodels 0.8.0
oauth2client 2.2.0
httplib2 0.9.2
python-dateutil 2.6.0
argparse 1.4.0
six 1.10.0
PyYAML 3.12
wrapt 1.10.10
crcmod 1.7
google-api-python-client 1.6.2
python-json-logger 0.1.7
gcloud 0.18.3
subprocess32 3.2.7 (removed when running in Python 3 training)
wheel 0.30.0a0
WebOb 1.7.2
Paste 2.0.3
tornado 4.5.1
grpcio 1.3.0
requests 2.13.0
webapp2 3.0.0b1
google-cloud-logging 1.0.0
The following Ubuntu packages:
curl
libcurl3-dev
wget
zip
unzip
git
vim
build-essential
ca-certificates
ca-certificates-java
pkg-config
rsync
libatlas-base-dev
liblapack-dev
libpng-dev
libjpeg-dev
gfortran
python3.5
python3-dev
python3-pip
python3-setuptools

python2.7
python-dev
python-setuptools
gdb
openjdk-8-jdk
openjdk-8-jre-headless
g++
zlib1g-dev
libio-all-perl
module-init-tools
libsnappy-dev
libyaml-0-2
python-opencv
March 13, 2018 April 5, 2018

Runtime version 1.5 supports TensorFlow 1.5.0 for CPU and GPU.

Python 3.5 is available for training and online prediction with runtime version 1.5. The Ubuntu packages for Python 3 (indicated in bold) are installed when running Python 3.

1.4 TensorFlow 1.4.0 and 1.4.1
The following PyPI packages:
xgboost 0.6a2
numpy 1.12.1
pandas 0.19.2
scipy 0.19.0
scikit-learn 0.18.1
sympy 1.0
statsmodels 0.8.0
oauth2client 2.2.0
httplib2 0.9.2
python-dateutil 2.6.0
argparse 1.4.0
six 1.10.0
PyYAML 3.12
wrapt 1.10.10
crcmod 1.7
google-api-python-client 1.6.2
python-json-logger 0.1.7
gcloud 0.18.3
subprocess32 3.2.7 (removed when running in Python 3 training)
wheel 0.30.0a0
WebOb 1.7.2
Paste 2.0.3
tornado 4.5.1
grpcio 1.3.0
requests 2.13.0
webapp2 3.0.0b1
google-cloud-logging 1.0.0
The following Ubuntu packages:
curl
libcurl3-dev
wget
zip
unzip
git
vim
build-essential
ca-certificates
ca-certificates-java
pkg-config
rsync
libatlas-base-dev
liblapack-dev
libpng-dev
libjpeg-dev
gfortran
python3.5
python3-dev
python3-pip
python3-setuptools

python2.7
python-dev
python-setuptools
gdb
openjdk-8-jdk
openjdk-8-jre-headless
g++
zlib1g-dev
libio-all-perl
module-init-tools
libsnappy-dev
libyaml-0-2
python-opencv
December 11, 2017 April 5, 2018

Runtime version 1.4 uses TensorFlow 1.4.0 for online prediction, and 1.4.1 for batch prediction and training.

Python 3.5 is available for training and online prediction with runtime version 1.4. The Ubuntu packages for Python 3 (indicated in bold) are installed when running Python 3.

1.2 TensorFlow 1.2.0
The following PyPI packages:
numpy 1.12.1
pandas 0.19.2
scipy 0.19.0
scikit-learn 0.18.1
sympy 1.0
statsmodels 0.8.0
oauth2client 2.2.0
httplib2 0.9.2
python-dateutil 2.6.0
argparse 1.4.0
six 1.10.0
PyYAML 3.12
wrapt 1.10.10
crcmod 1.7
google-api-python-client 1.6.2
python-json-logger 0.1.7
gcloud 0.18.3
subprocess32 3.2.7
wheel 0.30.0a0
WebOb 1.7.2
Paste 2.0.3
tornado 4.5.1
grpcio 1.3.0
requests 2.13.0
webapp2 3.0.0b1
google-cloud-logging 1.0.0
The following Ubuntu packages:
curl
libcurl3-dev
wget
zip
unzip
git
vim
build-essential
ca-certificates
ca-certificates-java
pkg-config
rsync
libatlas-base-dev
liblapack-dev
gfortran
python2.7
python-dev
python-setuptools
gdb
openjdk-8-jdk
openjdk-8-jre-headless
g++
zlib1g-dev
libio-all-perl
module-init-tools
libsnappy-dev
libyaml-0-2
python-opencv
June 27, 2017

Runtime version 1.2 uses an Ubuntu 16.04 OS base image instead of the Debian Jessie version used by 1.0.

1.0 TensorFlow 1.0.1
The following PyPI packages:
numpy 1.12.1
pandas 0.17.1
scipy 0.17.0
scikit-learn 0.17.0
sympy 0.7.6.1
statsmodels 0.6.1
oauth2client 2.2.0
httplib2 0.9.2
python-dateutil 2.5.0
argparse 1.2.1
six 1.10.0
PyYAML 3.11
wrapt 1.10.8
crcmod 1.7
google-api-python-client 1.5.1
python-json-logger 0.1.5
gcloud 0.18.1
subprocess32 3.2.7
wheel 0.30.0a0
WebOb 1.6.2
Paste 2.0.3
tornado 4.3
grpcio 1.0.1
requests 2.9.1
webapp2 3.0.0b1
google-cloud-logging 0.22.0
The following Debian packages:
curl
libcurl3-dev
wget
zip
unzip
git
vim
build-essential
ca-certificates
pkg-config
rsync
libatlas-base-dev
liblapack-dev
gfortran
python2.7
python-dev
python-setuptools
gdb
openjdk-8-jdk
openjdk-8-jre-headless
g++
zlib1g-dev
libio-all-perl
module-init-tools
libyaml-0-2
python-opencv
March 8, 2017

This is the default Cloud ML Engine runtime version currently supported by the v1 Cloud Machine Learning Engine API.

Support for older runtime versions

Older major and minor runtime versions are supported for a specific period of time after the new version is released. You can continue to train models and get predictions using an older runtime version during this period. Runtime versions are not supported beyond that window.

Months after runtime is superseded Can train models and serve predictions
0-12 Yes
12+ No

Support for GPUs

GPU-enabled machines come pre-installed with tensorflow-gpu, the Tensorflow Python package with GPU support.

Other machines come pre-installed with the regular tensorflow package instead.

What's next

Send feedback about...

Cloud ML Engine for TensorFlow