GitHub user witgo reopened a pull request:

    https://github.com/apache/spark/pull/2388

    [WIP][SPARK-1405][MLLIB] topic modeling on Graphx

    This PR relies on  #2631
    
    - [X] Topic de-duplication
    - [X] Support  100000 topics
    - [X] Asymmetric Dirichlet priors
    - [ ] Add the documentation
    - [X] Add infer interface
    - [X] Add unit tests
    - [X] Add the performance test 
    - [X] Optimizing the infer interface performance
    - [ ] Verifying the correctness of the algorithm
    
    
    The performance test:
    
    `2000` topics:
    
    Item | value
    ------------ | -------------
    The cluster resource | 36 executors(36 cores, 216g memory)
    The corpus size | 253064 document, 29696335 words
    The number of iterations | `105`
    The number of distinct term |  75496
    The number of topics |  `2000`
    alpha | 0.01
    beta | 0.01
    The running time |  37.1 minutes
    
    `10000` topics:
    
    Item | value
    ------------ | -------------
    The cluster resource | 36 executors(36 cores, 216g memory)
    The corpus size | 253064 document, 29696335 words
    The number of iterations | `105`
    The number of distinct term |  75496
    The number of topics |  `10000`
    alpha | 0.01
    beta | 0.01
    The running time |  49 minutes
    
    
    `100000` topics:
    
    Item | value
    ------------ | -------------
    The cluster resource | 36 executors(36 cores, 216g memory)
    The corpus size | 253064 document, 29696335 words
    The number of iterations | `105`
    The number of distinct term |  75496
    The number of topics |  `100000`
    alpha | 0.1
    beta | 0.01
    The running time |  268.9 minutes
    
    conf/spark-defaults.conf:
    ```
    spark.akka.frameSize   20
    spark.executor.instances 36
    spark.rdd.compress true
    spark.executor.memory   6g
    spark.default.parallelism  72
    spark.broadcast.blockSize  8192
    spark.storage.memoryFraction 0.4
    spark.serializer org.apache.spark.serializer.KryoSerializer
    spark.kryo.registrator 
org.apache.spark.mllib.feature.TopicModelingKryoRegistrator
    ```

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/witgo/spark graphx_lda

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/spark/pull/2388.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #2388
    
----
commit ca8e6f296a2f7ed674dd3a5cde49d4301d3d6d14
Author: GuoQiang Li <[email protected]>
Date:   2014-10-08T08:10:12Z

    topic modeling on Graphx

----


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to