-
Notifications
You must be signed in to change notification settings - Fork 20
Update DockerfilePytorch and fix CICD files #7
Conversation
zigzagcai
commented
Aug 29, 2022
•
edited
Loading
edited
- add spark lib in pytorch env
- add recsys support (lightgbm, xgboost, transformers)
- base image change from oneapi-aikit to ubuntu
- fix cicd scripts
@zigzagcai , I successfully built my docker using this new Dockerfile with one additional argument as below: I think we'd better add this info to README to guide users who also needs proxy when building docker, below is my BKM
|
ENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64 | ||
ENV SPARK_HOME /home/spark-3.2.1-bin-hadoop3.2 | ||
ENV PYTHONPATH $SPARK_HOME/python/:$PYTHONPATH | ||
ENV PYTHONPATH $SPARK_HOME/python/lib/py4j-0.10.9.3-src.zip:$PYTHONPATH |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I realized you only installed spark but no hadoop, which leads to we can only use local filesystem instead of HDFS as backend, personally I think it should be OK for now, but we may see this as a potential issue if we need to do distributed data processing.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay, Thanks.
Thanks. README.md has been updated. |
@Jian-Zhang @xuechendi Please help to merge this PR, so that @csdingbin can trigger DLRM CICD tests with the updated Dockerfile. |
Add tests for recdp and add built mvn to avoid extra building process
…model merge (#375) * change save path * simplify test code * rename path * add * add * add * add * add * bug fix * add * add * add * add * add * restore * add * add * add * add * Dtuner models (#7) * bug fix * add * add * update readme * delete * update test scripts * update * support direct eval after merging model * update * update * ssf load previous config * automatic fill deltaargs * update config * copy code to merged dir * add test * update readme for merge model * bug fix * check the code file existence * update readme * update model name list * refine test scripts * bug fix * bug fix * allow tokenizer to be None * bug fix * bug fix * fix import * update path * move merge testing scripts