Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67083534
[Test build #545 has
started](https://amplab.cs.berkeley.edu/jenkins/job/NewSparkPullRequestBuilder/545/consoleFull)
for PR 1290 at commit
Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67083539
The test logs have expired...rerunning
---
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
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67087128
@witgo @ankurdave Thank you for suggestion! I merged #3677 into the ANN
code and then I run 100 iterations on mnist (60K instances). The running time
improved only 1%.
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67087246
@jkbradley Thank you!
---
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
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67089450
[Test build #545 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/NewSparkPullRequestBuilder/545/consoleFull)
for PR 1290 at commit
Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67089972
It looks like the error is:
```
[error]
/home/jenkins/workspace/NewSparkPullRequestBuilder/mllib/src/test/scala/org/apache/spark/mllib/ann/ANNSuite.scala:21:
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67091484
@jkbradley This PR references an old MLlib that had this class
`LocalSparkContext`. It was substituted with `MLlibTestSparkContext` in one of
the latest releases. This
Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-67092852
Yes, Jenkins will test against the master branch, so I'd recommend merging
with master (or rebasing if the merge is messy).
---
If your project is set up for it, you
Github user witgo commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-66876255
@ankurdave
#3677 should be able to help improve this PR performance
---
If your project is set up for it, you can reply to this email and have your
reply appear on
Github user witgo commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-66876521
@avulanov I will submit a new PR about `AdaDelta` and `AdaGrad` in the next
week
It should be able to use in this PR.
---
If your project is set up for it, you can
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-66824545
@mengxr Could you suggest why the build has failed?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If
Github user bgreeven commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r21603916
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,239 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-66454787
[Test build #24310 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/24310/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-66461883
[Test build #24310 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/24310/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-66461896
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user ZhangBanger commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r21571009
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,239 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-65880449
@jkbradley @manishamde I did performance comparisons with multinomial
regression and posted them here:
https://github.com/apache/spark/pull/1379#issuecomment-65879536.
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-65731474
I've submitted a feature request to Spark Jira regarding the ANN-based
classifier: https://issues.apache.org/jira/browse/SPARK-4752. It is implemented
already and I
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-65264353
Just in case, I've tested the ANN without the hidden layer and it seems to
work as multinomial regression, though with (one-half) squared-error cost
function and
Github user Lewuathe commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-64176809
@avulanov What type of resource manager did you use for checking
performance?
I'm trying to look into the scalability how performance will increase
depending of the
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-64283686
@manishamde Thanks for the useful references! It seems that model
parallelization for ANN is a challenging problem. I asked this question to few
presenters on the
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-64285217
@jkbradley Thank you for good explanation! As for (1), single threaded ANN
implemented in C++ shows similar accuracy. (2) seems like some additional
coding needs to be
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-64285419
@Lewuathe Ganglia is used to monitor the cluster performance and it shows
that cores are busy all the way and memory underutilized. No resource manager
is used.
---
Github user witgo commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r20589805
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache Software
Github user jkbradley commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r20591105
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user bgreeven commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r20620263
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-63371703
@jkbradley (a) I found few papers with timing results, however with it was
with SVM. Moreover, the results differ from minutes to hours. Could you suggest
a paper with
Github user manishamde commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-63375342
@avulanov Thanks for conducting the experiments. Could you plot graphs for
the experiments that you conducted with changing number of features and number
of machines.
Github user manishamde commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-63375573
I found this reference recently about Netflix's distributed implementation
of neural nets that could be relevant for MLlib.
Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-63401446
@avulanov I'll look around for papers which might allow for comparisons;
I'm not sure offhand.
For experiments, I agree with @manishamde may getting
Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-63196649
@avulanov Thanks for running tests! It would be great if we could
calibrate tests somehow. Some options would be: (a) trying to reproduce tests
in a published paper
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62783988
@debasish83 I agree, such setup will not fit into the memory of a singe
node. Are you talking about recommender system? As far as I know, they are
usually addressed by
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62823148
@mengxr I've performed a test on mnist8m with our implementation of ANN as
you suggested. I used a 5-node cluster. Each node has Xeon 3.3GHz 4 cores with
16GB RAM. My
Github user debasish83 commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62824356
@avulanov I meant first sparse auto encoder with 1 layer and followed by a
deep net...these huge nets are mostly for automated feature extraction...of
course we can
Github user jkbradley commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r20181004
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,223 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62623318
@bgreeven @avulanov About supporting various optimizer, updater, gradient,
and error function options, I vote for keeping them as parameters, rather than
having
Github user witgo commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62254979
I agree with what @debasish83 said. We should find a suitable solution to
weight matrix distributed storage.
---
If your project is set up for it, you can reply to this
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62187026
@witgo Thank you for your suggestion! Could you elaborate how als algorithm
design could be used?
---
If your project is set up for it, you can reply to this email and
Github user debasish83 commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62190595
For matrix factorization we have user x product sparse matrix...You can
think of this sparse matrix as the feature matrix for ANN...Now consider two
matrices H1 and
Github user debasish83 commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62191470
f is neural activation...it can be tanh or sigmoid function (they are
non-convex, nonlinear) , LRU units (max is convex)...in this PR
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62195511
I think I am missing something from your description. In ANN we need to
compute K+1 weight matrices: W1(NxH1), W2(H1xH2), ... WK(HKxM), where N is the
number of inputs
Github user debasish83 commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62197587
I am considering 1 hidden layer ANN right now...For multiple layers I have
not thought on it since to scale it (if you want the model to be distributed) I
don't think
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62199347
I think before doing this we need to consider the benefits versus the
current implementation with loops. I can name one - the readability of code
will be better with
Github user debasish83 commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62247730
Our sparse data very often has ~10M users and 1M features...for 1 hidden
layer net with 10K nodes, we need 10M x 10K + 1M x 10K doubles...if the model
is not
Github user witgo commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62085010
We should use matrix to calculate the forward propagation ,back propagation
see
http://deeplearning.stanford.edu/wiki/index.php/Neural_Network_Vectorization
---
If
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62086344
@witgo I have implemented ANN in matrix form for Spark with breeze a while
ago: https://github.com/avulanov/spark/tree/neuralnetwork. We tested it with
@bgreeven and it
Github user witgo commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-62089562
We can not use existing Gradient classes,Let the whole iterative process is
completed in the form of matrix calculation.Moreover We can use the als
algorithm design, cut
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19922521
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,223 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user bgreeven commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61749598
Let's discuss a bit more about making the optimiser, updater, gradient, and
error function customizable.
Notice that for the current LBFGS algorithm, the error
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61757505
I think that these 3 parameters should be somehow bound otherwise one can
plug a gradient with vector length that does not correspond to the ANN size. We
could provide
Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61757723
With regards to performance tests, I am going to use mnist8m
(http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/multiclass.html) as
suggested by @mengxr. I have
Github user bgreeven commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19722711
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user bgreeven commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19723239
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,223 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user bgreeven commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19733377
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user bgreeven commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19733442
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61480481
[Test build #22818 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22818/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61481496
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61499579
**[Test build #22818 timed
out](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22818/consoleFull)**
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61499597
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61584664
[Test build #22851 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22851/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61590143
[Test build #22851 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22851/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61590149
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19717255
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user manishamde commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61433892
@bgreeven I haven't studied the implementation details yet but I had a
question about the API. I realize that RDD[(Vector, Vector)] is a more general
structure for
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19717451
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,223 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19717692
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19717699
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user manishamde commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-61436212
@bgreeven Another general suggestion: consider adding logging to the code.
It goes a long way in debugging errors and get statuses on long running job.
Check
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19717937
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19717985
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/ann/ArtificialNeuralNetwork.scala
---
@@ -0,0 +1,528 @@
+/*
+ * Licensed to the Apache
Github user avulanov commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19719009
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,223 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user manishamde commented on a diff in the pull request:
https://github.com/apache/spark/pull/1290#discussion_r19720701
--- Diff: docs/mllib-ann.md ---
@@ -0,0 +1,223 @@
+---
+layout: global
+title: Artificial Neural Networks - MLlib
+displayTitle: a
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-60742124
[Test build #22359 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22359/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-60749017
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-60749010
[Test build #22359 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22359/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-60844377
[Test build #22385 has
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22385/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-60853271
[Test build #22385 has
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/22385/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-60853279
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-57076438
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20930/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-57077703
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-57077701
[QA tests have
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20930/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56952211
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20857/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56957127
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56957114
[QA tests have
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20857/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56962296
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20858/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56971412
[QA tests have
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20858/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56971423
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56789789
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20793/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56790021
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56790721
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20795/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56791029
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56791103
[QA tests have
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20795/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56791106
Test FAILed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56791275
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20797/consoleFull)
for PR 1290 at commit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56796798
[QA tests have
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20793/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56796805
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56798291
[QA tests have
finished](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20797/consoleFull)
for PR 1290 at commit
Github user AmplabJenkins commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56798296
Test PASSed.
Refer to this link for build results (access rights to CI server needed):
Github user bgreeven commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56343909
Changed optimiser to LBFGS. Works much faster, but has the disadvantage
(due to the increased convergence speed per iteration) that it also starts to
exhibit
Github user SparkQA commented on the pull request:
https://github.com/apache/spark/pull/1290#issuecomment-56343891
[QA tests have
started](https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/20650/consoleFull)
for PR 1290 at commit
1 - 100 of 177 matches
Mail list logo
