moveforever opened a new pull request #18411:
URL: https://github.com/apache/incubator-mxnet/pull/18411


   ## Description ##
   (Brief description on what this PR is about)
   In recommendation, there appear many features: 
   - dense features, for instance, ctr, item exposure, item click, item statis 
features, etc.
   - one-hot categorical features, for instance, user's network, user's gender, 
item_id to score
   - mulit-hot categorical features, for instance, user profiles which 
describes interest, and there exists much interests for a user. So the length 
of the multi-hot categorical features is variable.
   The sample can be as followed.
   
![image](https://user-images.githubusercontent.com/5248288/82135079-33513180-9831-11ea-92a4-83b6fe0d8aba.png)
   The row is splited by ^A(\001). There are five columns. The first column is 
label, supporting multi labels. And the second column is dense feature, and the 
third column is categorical feature, and the the fourth column is multi-hot 
categorical feature which can be splited by comma , and the fifth column is 
sparse feature which is support wide input for google wide and deep model. 
   
   The sample contains many formated features which are storaged by csv or 
sparse. So it is complicated. To high performance parse the sample, i implent 
c++ DataIter. Train at gpu will not be hold back by cpu parsing sample.
   
   In the meantimes, part code of dmlc-core is mended where is in 
https://github.com/moveforever/dmlc-core-recomm_data_iter/commit/7aa77953799d0ac152ddc5d3d32bf36988bbac36
   Reference:https://github.com/apache/incubator-mxnet/issues/18346 
   ## Checklist ##
   ### Essentials ###
   Please feel free to remove inapplicable items for your PR.
   - [ ] The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to 
the relevant [JIRA issue](https://issues.apache.org/jira/projects/MXNET/issues) 
created (except PRs with tiny changes)
   - [ ] Changes are complete (i.e. I finished coding on this PR)
   - [ ] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [ ] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, README.md is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - Check the API doc at 
https://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
   - [ ] To the best of my knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change
   
   ### Changes ###
   - [ ] Feature1, tests, (and when applicable, API doc)
   - [ ] Feature2, tests, (and when applicable, API doc)
   
   ## Comments ##
   - If this change is a backward incompatible change, why must this change be 
made.
   - Interesting edge cases to note here
   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to