Skip to content

Latest commit

 

History

History
96 lines (61 loc) · 2.85 KB

changelog.md

File metadata and controls

96 lines (61 loc) · 2.85 KB

23.09.16

  • upgrade base to tensorflow/tensorflow:2.13.0-gpu
  • retire torch packages
  • introduce dask[complete] and seaborn[stats]
  • retire extra tf packages
  • retire langchain

23.05.20

  • switch to tensorflow 2.12.0-gpu from 2.6.1-gpu
  • comment rasa as a core dependency

21.11.30

  • switch to tensorflow 2.6.1-gpu from 2.6.1-gpu-jupyter

21.11.28

21.11.18

  • upgrade to tensorflow 2.7.0-gpu
  • remove hack to install python 3.8

21.04.26

  • restore docker builder from kaniko

21.04.25

  • add sci-spacy and spacy
  • remove RUN python3.8 -m pip install -U https://github.com/PythonOT/POT/archive/master.zip -- user

21.02.07.2

  • correctly install python3.8 dependencies

21.02.07

  • remove tf-models-nightly. This will significantly reduce the image size
  • default to python3.8

21.02.06

  • add tensorflow-hub, tensorflow-text, tensorflow-addons, datashader, holoviews
  • change base to tensorflow/tensorflow:2.4.1-gpu

21.01.17

20.12.25

  • remove t5 package introduced in version 20.11.12. With this change, the image size should reduce significantly. If there is a need to restore it in the future, include the following line in the Dockerfile

      RUN python3 -m pip install git+https://github.com/kineticengines/text-to-text-transfer-transformer.git@mirror-main --use-feature=2020-resolver --user

  • change base image from 2.4.0rc0-gpu to the stable release2.4.0-gpu

20.11.22

  • change base from tf-nightly-gpu to 2.4.0rc0-gpu

  • introduce t5 package for text-to-text model experiments. t5 install torch. this may cause conflicts in tensorboard installation. restore fix from previous commit <920c5a71158bb310394a0cef197433067612f566>

  • remove java/spark comments in Dockerfile. in the future is these packages are needed agains, add the following line in the dockerfile'

    RUN apt-get update -q -y && apt-get -q -y install openjdk-11-jdk && echo "JAVA_HOME=/usr/lib/jvm/java-11-openjdk-amd64" | tee -a /etc/environment && \
    source /etc/environment

ENV SPARK_HOME=/opt/spark
# # Append spark pato to system paths
ENV PATH=$PATH:$SPARK_HOME/bin:$SPARK_HOME/sbin

RUN curl https://downloads.apache.org/spark/spark-3.0.1/spark-3.0.1-bin-hadoop3.2.tgz > spark-3.0.1-bin-hadoop3.2.tgz && \
  tar -xzf spark-3.0.1-bin-hadoop3.2.tgz && mv spark-3.0.1-bin-hadoop3.2 spark && mv spark /opt/

20.11.15

  • introduce docker best practises;
  • deprecate latest tag
  • deploy using calver for tagging new images
  • add kineticengines as default container user