You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by cm...@apache.org on 2018/10/05 00:45:56 UTC
[incubator-mxnet] branch master updated (e41faf0 -> 931555c)
This is an automated email from the ASF dual-hosted git repository.
cmeier pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from e41faf0 [MXNET-908] Speed up travis builds to avoid timeouts (#12706)
add 6568864 initial empty commit
add 1b1dcf9 MXNet.jl generated files.
add 3e0abce MXNet.jl regenerated files.
add e25bf7e basic API testing, create NDArray
add 11029be a placeholder README
add d7f011c move gc to handle finalizer
add 28531c8 define binary dn array functions
add e426493 copy ndarray to cpu for debuggin
add 0062c9d define unary ndarray functions
add 0ee41be use a prefix instead of submodule for dynamically imported functions
add f3184ad a more general way of defining mx imported functions
add e3a4119 only the generic way of defining function is enough
add 5ed258f common array-like interface functions
add 59dcb7e inplace operator for +
add 678f6f2 basic test for ndarray
add 525c5b2 comment out release testing until v0.4 is released
add e0a0bbe fix copy from and to CPU
add 6294c93 ndarray plus and minus unit-tests
add d6a1367 use more meaningful name for sub! and add!
add d33dff8 assignment operator for NDArray
add b2acdc7 test ndarray elem-wise mul
add 4667583 unittest ndarray div
add 12ca85a require julia v0.4 to run
add 65cc788 basic symbol type
add 4170e37 possibility to add dynamic doc string to imported functions
add 6490928 try setting up travis CI
add 573eedd fix travis script
add b1ca062 more fix for travis script
add 7a66ef4 enable osx build on travis
add b22c83f fix build script
add f9b88b0 fix typo
add bdd798a properly switch directory
add 7713589 fix sed option
add 8950df5 disable openmp for osx build
add 694f0bb add badges
add efb0b90 clean up built script
add 7626fe7 list symbol creator functions
add 627a8e5 name manager
add 5fea004 symbol constructor
add ce68e59 symbol basic test
add 8701e0a symbol test_internal
add 320706f symbol test_compose
add 631a965 test infer shape
add fca4789 test infer_shape with error
add c05e6a5 try out julia v0.4 precompilation
add dd1922d executor
add 0a04f25 executor bind forward
add 2074d12 executor backward
add 6dc48bb simple_bind
add 9dd26e5 random module
add 357a105 default initializers
add f3e4e95 non-unicode name for initializer parameters
add 7f35875 make clear the meaning of row-major vs. column-major in array conversion
add 23c14fc check writable bit in NDArrays
add 7481dab untested io.jl
add 956a220 unittest MNIST data provider
add 3113256 kvstore init / push / pull
add 8e4102f kvstore test aggregration
add 48a2b20 eval metric
add dd02638 shape should be translated between Julia<->C in infer_shape
add d061a18 optimizer and estimator
add 8a0a43b fix initializers
add 92ec499 fix executor
add 8c0e465 bug fix
add 614de3a mnist mlp example
add 61fa00c lenet example (not tested)
add 383fd98 fix kvstore closure
add 7f8612e cifar10 example (pushing to GPU server for testing)
add edf0871 working cifar10 example
add aa18e98 basic doc building setup
add ff055b5 test rtd hook
add 4a25b00 add rtd badge
add 34a1c89 use rtd theme
add 4eff47d fix some doc markdown formatting
add d13ddc6 clean up Context
add 70579c4 use light-weighted mkdocs for documentation instead
add 0257055 one-sentence readme.
add 1eb7ebc use osx.mk to simplify travis OSX build
add 9223721 clean up CI build script
add 2c945ea docstring for imported symbols
add 1da331c installation guide (or what we wanted to be...)
add 684bad2 part of overview
add df6613d shape inference doc
add 4ac5e7b allow keyword arguments in bind.
add af56456 part of the doc on NDArrays
add cf9a06a getindex operator for NDArray
add 3a26544 ndarray doc
add 0c3d066 API variable -> Variable; group -> Group
add 9665cb1 update doc: variable -> Variable
add 210aba8 add libmxnet.dll into library search list
add 478568c doc for mnist
add d15bc77 add mx.chain macro
add e9088f5 rename init.jl -> base.jl
add bed3a02 refactor optimizer options
add a75a8ef simplify mnist mlp example
add 43b2486 clean up lenet example
add 02e4af3 hygiene in chain macro
add 1fb4946 use @chain in lenet example.
add 3548a5f lenet tutorial
add 7c43f2c update readme.
add 3b43bcb first version of deps/build.jl that is not working b/c opencv
add e77fdd0 default building without opencv.
add 263e602 update install guide
add 7934f67 add release notes for v0.0.1
add 0a32996 fix build script (don't exit Julia)
add 48f4230 update news for bugfix v0.0.2
add 119fac1 update kvstore doc.
add d738fba NDArray save and load API
add 4e5e152 gradient clipping
add 9d138da save and load for symbol
add 30f89a8 unit test for symbol save and load
add 7e86c8f improve defstruct macro
add 00bd457 callback interface
add 06879c6 fix estimator API
add f551c91 add speedometer to cifar10
add cb067c3 fix speedometer
add f7d7ce3 save checkpoints in callbacks
add 5ea6b8c array iterator and (part of) the unittest
add b4cc05c more test on array data iterator
add 2c9d9ae more tests on array data provider
add eb84402 test shuffling of array data provider
add 1538340 rename estimator -> model
add dc8ca18 fix README example
add 2a2b425 Merge pull request #6 from yeesian/patch-1
add 6d87215 Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add 28791db use joinpath in README
add 1baeabc expose init_model API
add 7fa1511 model predict API and demo
add bb4620e model checkpoint loading
add f8a01bf allow callbacks in predict
add 0f632cc add pretrained Inception model
add 5bcd859 add example using pre-trained model to do prediction.
add 5a80bf3 fix #8, more informative error when libmxnet not found
add 8bb82e4 use sphinx as documentation system
add bcc90ef switch from mkdocs to sphinx
add 8d56a9f a dead simple embedded sphinx doc system (cf #3)
add f1f869f API document for NDArrays
add ac39fd6 libmxnet API for ndarrays
add e7ec661 automatic build API for symbols
add 9c87c61 fix mal-formatted param type in sphinx syntax
add 091c234 add title to symbol doc page
add 54e621e prepare for v0.0.3
add cadee3a better type annotation in docs.
add 1b62ddf more docs.
add d19f28f docs for built-in callbacks
add c051c2a fixed joinpath problem in example in README.md
add de566eb Merge pull request #9 from cbecker/master
add 0132f19 added ADAM SGD
add c29e18a Merge pull request #10 from cbecker/master
add 6e0c5c7 some doc for symbol
add 762c45e refactoring and naming convention dmlc/mxnet#407
add a2beae8 refactoring of optimizer
add fe36340 refactor adam optimizer
add 1f39e1d more learning rate scheduler
add e821f95 update equation in sphinx doc
add 04bde14 fix accidental notebook changes
add 1dc944e first step of refining the data provider API.
add 49a5acb import doc of built-in data provider.
add 76fee6b refine mxdataiter api
add 5e03d27 fix unittest for new data IO api
add ceaa3eb fix mlp example to use the updated data io API
add 993be59 simplify model a bit with data IO api
add f30aedc part of the executor API doc
add 388e91d fix mnist doc links
add 319c5f3 rename Softmax -> SoftmaxOutput (mxnet/mxnet#434)
add 6389094 update doc Softmax -> SoftmaxOutput (dmlc/mxnet#434)
add c1d663d run mlp.jl on travis CI
add 922fc85 Reenable 0.4 testing.
add 4acbce9 Add coverage.
add ebb573a Merge pull request #13 from invenia/coverage
add 3d447a4 Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add b1e882b fix env variable for detecting Travis CI
add f823510 only run 1 epoch for MNIST on travis CI
add 23378f1 fix error due to API changes of subarray in Julia nightly
add d26c36b a specific mlp script for Travis CI regression test.
add 1786bdb wrap mnist test into a module
add 3bf4364 fix travis test error
add d0dadad pull-back only when update_on_kvstore
add 75b3e68 SVMLight Data Provider
add f6cbba1 add description of SVMLight Provider
add 0f22cfa add readme for plugins
add 726badd update doc for updated data iter API
add 7ed591d add libatlas-base for cblas
add 7e89a78 Merge pull request #17 from vchuravy/vc/travis
add ee855ec fix cifar10 example
add 8486b97 Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add 637bf0c avoid calling mxdataiter reset
add ee0583f remove unused variable
add 315c1a4 simplify a bit io test
add 722eb36 fix array data iter provide size not properly reported (see #14)
add 189e755 prepare for v0.0.4
add 76261c7 refactoring: Symbol -> Node
add 955b6d2 refactor: Node -> SymbolicNode
add aab8784 upate doc for refactoring
add c13f88f add nn-factory
add 3bc475a add doc for nn factory
add 41c20ae fix travis CI build error due to refactoring
add 3e8627d document clean up
add be8cbda more test for IO
add b49dc62 update doc (Symbol -> SymbolicNode)
add 591871c add doc for metrics
add 2b85c9c share NDArray data with Julia Array.
add 2055727 nd_as_jl macro
add ea90b55 use nd_as_jl to re-write accuracy
add b5c0d96 reformultate accuracy with multi_output in mind
add 8322b63 Calculate accuracy based on size of pred.
add c300971 remove sub from accuracy
add d51d2af add comments for accuracy and rebuild docs
add 7947313 Merge pull request #23 from vchuravy/vc/multiout_acc
add 4362525 add nd_as_jl unit test
add e884692 update ipython notebook BGR -> RGB
add 3226713 complete doc for symbolic-node
add f5f2777 Adds Mean Squared Error evaluation metric
add 94e17a6 Merge pull request #25 from Andy-P/pull-request/f5f27779
add 0b538d1 make notes on MSE metrics
add 645091c char-ltsm temp commit
add 703d78f char-ltsm example, not tested
add eea7d2c typo ltsm -> lstm
add e35b929 fix macro error reporting
add 5a9d56a fix some errors
add 487d5e8 fix kvstore refactoring bug
add e4b6306 fix databatch
add 6ddbecd add speedometer
add 00c7bff neg-log-likilihood eval metric
add d685f77 tweak training log formatting
add 69cc430 reset! for NLL calculation
add 0e1515d fix lstm learning rate
add c5c1e05 basic network visualization
add e0857a2 add visualization doc
add c4abb97 fix visualization bugs
add c3bee16 Merge branch 'master' of github.com:dmlc/MXNet.jl
add 6193ef4 add lr decay to lstm training
add 0c3f32d tweak learning rate
add b078fec lstm sampler (not tested)
add b624089 Merge branch 'master' of github.com:dmlc/MXNet.jl
add 0f1fbaa fix sampler
add f915865 use adam optimizer
add a6886c2 copy states over for sampler
add a153c3c Merge branch 'master' of github.com:dmlc/MXNet.jl
add db7ae38 fix LSTM sampler
add 8829912 add basic readme for lstm example
add 5da45f3 use a more sensible default argument in favor of #24
add 3bc4fb5 fix path in char-lstm example outputs
add 0e6402d add simple visualization example of char-lstm
add a392e04 tweak visualization
add 379de2b part of LSTM tutorial
add 5b9df89 update README for LSTM example.
add 75c32cd Add gitter link to dmlc/mxnet
add a9242f4 Merge branch 'master' of github.com:dmlc/MXNet.jl
add b2b903c Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add c804be9 char-LSTM tutorial
add b595394 call visualization in test
add f099670 prepare for v0.0.5
add e8ab01d tweak visualization style.
add 37b48f1 fix the sample code in README
add a35c709 Merge pull request #27 from bicycle1885/patch-1
add c2d4b1e Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add aea096f use a smaller network for visualization.
add 4f09440 implements variants of Xaiver
add 56bd4f7 Merge pull request #28 from vchuravy/vc/fix_xavier
add 89cf70d update doc
add 1a935de rename Xaiver to Xavier
add 5dd8664 rebuild docs
add 0b38010 Merge pull request #30 from vchuravy/vc/rename
add 5b0e1ac remove grad_scale parameter from optimizers (duplicated with lr)
add 4fc66b4 update documents for optimizers
add 884c7f8 more arithmetic operations on symbolic nodes
add 7d8de6b fix API change: scalar_on_right -> on_left
add 62c9703 rework Xavier to be more flexible
add 6081fce xavier: rebuild documentation
add ea85774 Merge pull request #32 from vchuravy/vc/xavier
add 56f3eef update README
add e848e8d spelling :bee:
add 03eb95f Merge pull request #33 from yeesian/update-readme
add 2c72446 Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add 46257d2 basic interface for setting and getting attributes
add 05bcddf Merge pull request #35 from vchuravy/vc/symbol_attr
add e19ca8e add warning to discourage users from calling set_attr directly
add 266d946 attr interface for SymbolicNodes
add 247ec44 attrs: Doc rebuild
add 1f4bc8b Merge pull request #36 from vchuravy/vc/attr
add c66531f adds LearningRate.Inv from Mocha
add bb7e542 Merge pull request #37 from vchuravy/vc/lr_inv
add 74cc5ba Fixes usuage of floor
add 0d11ebc Merge pull request #42 from vchuravy/vc/fix_ops
add f094a11 add missing _init_default
add 2ff90c3 imports operators into the namespace to prevent issues
add fdf536c Merge pull request #46 from vchuravy/vc/ops
add 67c8bb2 prepare for v0.0.6 (c.f. #42)
add aa1b600 backport @defstruct from SimpleStructs.jl
add b513898 use download instead of wget
add 8d3a9a0 Merge pull request #48 from SimonDanisch/patch-1
add d612faf Accuracy metric added to epoch callbacks..
add a5c6693 Accuracy metric added to epoch callbacks.
add 49a92d1 Accuracy metric added to epoch callbacks.
add 9286bcc Merge pull request #50 from kasiabozek/acc_callback
add 71b4945 Fix compilation errors for julia 0.4.2
add 8e879d7 Fix compilation errors for julia 0.4.2
add 32c8954 Merge pull request #53 from BigEpsilon/patch-4
add 5ed54d8 try to fix build-script for installing cblas
add 0169981 try to use the built-in bindeps for building in travis CI
add f91c200 add Juliapkg badge
add dfe66f0 Merge branch 'master' of git://github.com/dmlc/MXNet.jl
add d8e5065 fix travis CI building script
add 7a7cfbe trying to fix travis CI gcc-4.8 thing
add 99e7e66 still trying to fix travis CI gcc issue...
add 3861edc clean up travis CI script
add 3a134e8 test whether bindeps works well on travis CI to install cblas
add ce57a39 test failed, travis CI is configured to disable sudo
add 8edb94b prepare for v0.0.7
add 4c52eb8 workaround for #55
add 724007a add notes about AWS GPUs
add c5cbf24 fix seg-fault due to upstream API change
add 175d399 workaround for jenkins CI
add f277cd7 revert last commit (not useful)
add 30b7142 fix one travis CI test error
add 1e4f9f0 fix error due to upstream API changes on symbolic node function
add 579b881 fix for breaking changes in Julia nightly
add ce3f6d3 another attempt to fix julia v0.5 compatability
add 938b236 update doc for initializer
add 06019e2 stricter type anotation for julia v0.5
add 1c707a2 fix typo
add ac709d1 prepare for v0.0.8
add 7a50357 MXNet exposes a transpose function
add fbcd12c Merge pull request #71 from oist/vc/imports
add f898837 add function to get debug_str, similar to python
add f7562e9 Merge pull request #73 from oist/vc/executor
add acd2a74 output TempSpace allocation
add 34efa24 Merge pull request #74 from oist/vc/temp_space
add 8b8c9c2 cleanup Accuracy
add 9182148 add average cross-entropy metric
add 3170c4e update documentation
add f6b7f31 Merge pull request #79 from oist/vc/accuracy
add 0b387f7 add further information to install.rst
add f26ad96 Update install.rst
add 79bfdbe Merge pull request #60 from ultradian/patch-1
add d85549b adds MultiACE or ACE per class
add 3bff587 Merge pull request #81 from oist/multi_aces
add 1a73dda Use openblas from Julia, instead of installing yet another BLAS library
add 0c5086a Simplify linking
add 7be95f6 Bundle generic cblas.h header for users that don't have libblas-dev installed already
add 0ec6dca Oh, it really does call itself `libmxnet.so` on non-linux platforms too
add 60c16e2 Cut down cblas.h to only the stuff we need, make this work on ILP64
add 7f445b9 Fix typo
add 9474561 More intelligent parallelized build
add 39786c5 Just remove parallel build, it's making Travis angry
add a2164ae Merge pull request #82 from staticfloat/sf/openblas
add b38e2cc update docs
add bbe3151 fix for upstream changes dmlc/mxnet#1857
add 4bf4037 check if their are predefined weights and use them
add 4c7a066 be more lenient in model loading
add 539a070 Merge pull request #86 from oist/vc/transfer_model
add 98b7cf2 add SymbolListAttr
add b0fc16f get idx that need to be frozen
add ce599a9 skip parameter update on frozen weights
add 9880bdf set GRAD_NOP for all frozen params
add 68180f7 ids derived from arg_names and param_names are not relatable.
add 5ba27eb Merge pull request #87 from oist/vc/finetuning
add deb021d (temporarily) disable Julia nightly test
add c948dc3 Added 7z alternative for mnist
add 69e9898 Merge pull request #90 from Mike43110/win/zip
add 013b88b Change where set_attr is called in _define_atomic_symbol_creator
add 2b0d825 Merge pull request #91 from Mike43110/list_attr
add 463f202 add support for ListAttrShallow
add e7cd135 Merge pull request #92 from vchuravy/vc/fixup
add baa9c2a revert #91
add 64a63f7 Deduplicate and generalise metric update!
add 0ea4369 Merge pull request #94 from oist/vc/generic_metric_update
add 3a013b8 fixup expr check for v0.5
add e6d07f0 fix deprications on v0.5
add 55f2a42 fix overwriting of inner function
add 48b723a fix ByteString deprication and deprications in tests
add afd9baa enable travis for 0.5
add d1bf894 Merge pull request #98 from oist/vc/0.5
add 9c9e762 Initial support for DType and TypeFlags
add a282586 add typed empty, ones, zeros
add c161fe1 make arithmetic work with DType and extend arithmetic tests on all types
add 2916a8f turn macro into helper function
add dc405e7 add type_infer
add 4a4ea61 make bind tests use isapprox and don't run division tests for Integer types
add a80ae98 Merge pull request #97 from oist/vc/ndarray
add affebe6 Exclude division test for Float16
add 9a827f7 fix blas_vendor for v0.5
add 1940a68 Merge pull request #102 from oist/vc/test
add 5e70cb9 start transitioning to Documenter.jl
add ce16bfe split import_ndarry functions up
add f09362c add mxnet.so as a dependency
add 411a08b simplify _import_ndarray
add 7b65547 implement _import_ndarray functions as macro
add 89e2c8e implement a function cache
add 5b2bd3e Refactor _import_atomic_symbol to be a macro
add 1ad9a1a transition context to see how it turns out
add ca949d7 get rid of the most of the old docs
add 504aec6 do some housekeeping and start moving things to markdown
add 733d46d setup api docs with autodoc
add 9113b21 rough first pass .rst -> .md
add d8de742 improve autogenerated documentation
add 597c736 convert rst to md with pandoc
add d7d4d90 finish mkdocs support
add 3dbec0f cleanup docs
add c74594f enable travis to build docs
add 85ec111 finalize mkdocs setup
add 8dad674 Merge pull request #106 from oist/vc/docs
add 789bb6b setup Documenter.jl
add 2c651ff fix documenter config bug
add f442c49 switch doc to new location.
add 79f51e6 fix URL to Windows building instruction (#111)
add b54620b Fix broken documentation link.
add 8058ff7 Merge pull request #112 from slundberg/patch-1
add a7b1c65 fix deprecations in build.jl
add 0408e3e fix String deprecations
add 3145973 Change back to dict comprehension syntax for v0.4
add 016efd6 improve docstrings
add 5088242 Don't override help for already defined methods
add 3bc93fe improve documentation formatting for SymbolicNodes
add b328f24 fix deprecation for call
add 2c36a5d nightly is v0.6
add a0a4a4e Merge pull request #115 from oist/vc/deprecations
add 63e6063 fix Inception URL #113
add 513c030 fix link in doc
add 3a7913f fix some formatting errors in docs
add 63772b8 fix typos in doc
add 50a2946 documentation cleanup
add 71eefbb Remove `` and :class:
add 8c22211 Merge pull request #116 from vchuravy/vc/doc
add 221611a Get rid of sub depwarn
add e672e42 import Compat.view
add 08ba2d1 Update REQUIRE with min version of Compat
add de77bce Merge pull request #117 from ranjanan/ranjanan-depwarn
add e3b7935 reorder docs for io.jl and bring them into the markdown age
add c126895 use `!!! note` syntax (this breaks documentation creation on v0.4)
add 3bbd866 build documentation on 0.5
add 84866f4 fix documentation being attached to the wrong binding:
add da6d9e4 Merge pull request #118 from oist/vc/io_doc
add 053ab42 fix overseen typo
add d79f9df Merge pull request #119 from oist/vc/typo
add 4403ef3 add MultiMetric and reorder metric docs
add 12a87e5 add one-dimensional ACE
add ab19db8 update docs for ACE
add 0621b13 Merge pull request #124 from vchuravy/vc/multi_metric
add e697960 add grad function to do autodiff
add a01a11b Merge pull request #125 from vchuravy/vc/symbol_grad
add 637e378 cleanup docs a bit more
add abc8ffa Merge pull request #127 from vchuravy/vc/clean_docs
add 1ba3896 update default dataset URL (fix test failing of #123)
add 3fac768 associate docs with method and not binding
add f0b758b move call docs into type docs
add 57aa885 Merge pull request #128 from vchuravy/vc/method_docs
add 6840119 handle kwargs for ndarray functions
add e0b7668 special case transpose
add f5e80af add special casing of transpose to the symbolic layer
add dc43bfe Merge pull request #123 from vchuravy/vc/ndarray_funcs
add ae1d45b preparing for v0.0.9
add 9e3a60a Positional arguments in symbolic API need type restriction. (#134)
add 8949dbb remove mistaken + in REQUIRE (#135)
add f06d994 fix special handling of dot for SymbolicNode (#123)
add 6e2c1ae add readme for developers
add ebdafe8 update libmxnet SHA
add 545332d keep the master branch to track libmxnet master
add 39bc643 make is_windows compatible with Julia v0.4
add ba2c999 fix build script
add 8f43961 prepare for v0.1.0
add 86362f3 update workflow for creating release
add 1a0b440 Make build.jl aware of other blas vendors besides OpenBLAS (#137)
add d4278ca remove unnecessary @compat call
add 1ff0975 replace pipe with pipeline
add 37e715a fix build script
add 9798c97 improve buildscript
add eb4a3fd fix build script.
add 44d659a Fixes deps/build.jl on Travis (#142)
add e218d85 fix NEWS.md for skipped v0.0.9
add 286d751 Add files via upload
add 45c1a6c Merge pull request #147 from mcreel/master
add 8453c0a Update regression-example.jl (#148)
add 6bd3951 try to automatically find CUDA
add 9059c00 Merge pull request #144 from vchuravy/vc/cuda
add 4664f3b Add some docs to `eachbatch` (#149)
add eee5ca7 fix missing wget #150
add 5880b77 Merge pull request #151 from dmlc/vc/fix_150
add d1eb318 Bug fix updating the ACE loss function. Included a test case. Also provided an example.
add c06b211 Merge pull request #156 from JockLawrie/pull-request/d1eb3183
add b81b26c New Optimizers (#159)
add 1ea44c9 Fix hygiene bug (#164)
add ab80048 ENH: added verbosity training option to control printouts (#162)
add 7a2432c ignore vscode project files
add e6b0508 ndarray ops
add d16b654 ndarray op test
add 03509ae temp commit, symbolic not passing due to ambiguity
add 623fbaf disambiguate nd op and sym op
add d718cfc fix API changes in random
add 590055b change examples for new symbolic calling conventions (nnvm, #146)
add 4d0aa87 Fixed visualize (#163)
add 2393d4a Fix scalar operations in NDArray (#165)
add 179daa5 Fix of bugs in nnvm branch (#169)
add 1b09fb4 Merge branch 'nnvm'
add 030990a BUG: fix two small mistakes with verbosity control (#168)
add 22a3c72 Drop support for v0.4 and Deprecations for v0.6 (#171)
add 2a66ea3 Update build.jl (#174)
add 3ad5819 update NEWS.md
add 8d044b1 refactor windows support
add 15c7052 Merge pull request #186 from dmlc/vc/windows
add b039771 Using WebRequest function to download the JSON file
add 11deaab Merge pull request #188 from bmharsha/Fix-Windows
add 6335eca fix vectorized abs deprecation
add d9f9551 don't include code from within a function
add a6968a6 deprecation of filter and Array creation
add a79d002 readall deprecation
add a476d3e MXNet.jl on v0.6 is currently subtlety broken due to broadcasting.
add ba3c9fb Merge pull request #187 from oist/vc/travis
add fecfcdd fix use of Symbol in symbolic-node.jl, fixes #189
add fb2a0d0 Merge pull request #190 from dmlc/vc/fixup
add cac5625 refactor sgd.jl to exploit multiple dispatch
add d0d7963 remove BaseTestNext and bump Compat
add d59b214 Merge pull request #192 from dmlc/vc/review_comments
add 0ee926e remove usr/setupenv.cmd because it is too invasive
add 79d80b6 Merge pull request #196 from dmlc/vc/fix_win
add 75060bc Fix for symbolic node (#199)
add f33f1dc get_name (#206)
add 17a5442 fix data URL error (#204)
add c58eb46 fix for issue #207 (#208)
add fa25ad6 metric: intro NMSE
add f23b0a9 doc: update CDN link of MathJax
add 96191ea metric: add test cases for NMSE
add ba97d93 docs: add an explanation for normalized MSE
add 09d8f27 Merge pull request #211 from iblis17/nmse
add cf7f32b Fixed data attribute
add c9eb71b switch to Plots.jl from pyplot (#217)
add 1781290 Update ACE metric (#220)
add 9474d95 provide an optional context in mx.load_checkpoint (#221)
add 46440e3 Added initializers for Spatial Transformer Networks (STN) layers (#223)
add 9dd3291 docs: update renamed `every_n_batch`
add fc2974f Merge pull request #225 from iblis17/docs
add 5c02fcb Check if key stride exists in the node info when producing graphviz output. Fixes #213
add 867b1c5 Merge pull request #226 from facundoq/master
add 31b95fb Update mnist.md (#227)
add 1be1b87 fix conflict with Images
add 7a6120a Merge pull request #228 from dmlc/vc/images_subtype
add bfa966a fixes bilinear initializer following approach in #34
add 8947ead adds test for bilinear initializer
add 44a6c36 Merge pull request #229 from dmlc/vc/bilinear_initializer
add ced3f73 fix pointer(NDArray) for Float64
add 86e3fa0 Merge pull request #231 from dmlc/vc/pointer
add 558adb7 add SeqMetric to apply different metrics for multi-output
add b4ece81 add NullMetric
add 3263b95 Merge pull request #230 from dmlc/vc/seqmetric
add d0253f1 XCode8 supports thread_local in C++
add 68b67e8 Merge pull request #234 from dmlc/vc/mac_travis
add 1697522 update MSE to be agnostic to dimensions.
add e12052c convert single outputs to array so that inference has it easier
add def129f convert eps to eltype of array
add c2044d5 adds THTT to enable metrics to support NDArray or Array
add 349d2a7 convert MSE to use NDArray as an example for async
add 4675ace Merge pull request #233 from dmlc/vc/metrics
add c4a2658 ensure that new submodules are initialized
add 9aac7e1 Update callback.jl (#235)
add 7a530fa Fixed SpatialTransformerNetwork loc_bias initializer. Added lenet-stn example. (#238)
add 06010b8 explicitly specify 7z (#239)
add 6cc9e99 add verbosity option to predict()
add 4cd5f3f Merge pull request #244 from rdeits/predict-verbosity
add 5d16305 explicitly specify 7z
add 02c7661 Merge pull request #246 from musm/patch-1
add 8918f1b update appveyor script (#195)
add efff124 Add appveyor badge
add 60f5306 improve GPU detection.
add 07777d6 switch to vc14
add 28344a4 actually copy the mxnet.dll to the right place on windows
add 96bdfa7 Prebuilt binaries only work on windows 64bit
add aae9114 Merge pull request #194 from dmlc/vc/gpu_detection
add 2ecf2c1 update NEWS.md
add b63fd9b Fix #242 (#247)
add 9c70922 Update CI URLs to point to new caching infrastructure
add c7032f0 Merge pull request #249 from staticfloat/ci_url
add 172af60 Use `Sys.CPU_CORES` not `nprocs()` when doing a parallel `make` (#250)
add 2b26bc2 update for Julia v0.6 (#248)
add 841e9ee update regression example for julia v0.6(#254)
add c965235 Add extra check on CUDA library detection
add 18547b8 Merge pull request #255 from joshbode/cudalib_check
add e7d1b62 Added cblas_[ds]tr[sm]m to cblas include file (#259)
add 10ad90b Added MXNET_COMMIT env variable for upstream commit/tag (#260)
add 0750c41 build: bump curr_win to 20170819, MXNet 0.11.0
add 778529c Merge pull request #268 from iblis17/win-prebuild
add 873715b cblas: import all func prototype into header
add 85a2d97 ndarray: add kInt8 and kInt64 into TypeFlag enum
add 8bd307f Merge pull request #270 from iblis17/nd-typeflag
add 4e5ab13 travis : set make -j4 and travis_wait
add f34643b Merge pull request #271 from iblis17/fix-travis
add d1bbcc4 base/build: using `Libdl.dlext` for searching lib
add b68ce5b Merge pull request #265 from iblis17/ext
add 0d54f4b build: force enable LAPACK when Julia's OpenBLAS is available
add 21ccc56 Merge pull request #269 from iblis17/build-lapack
add 23866bd build: fix macOS LAPACK build
add 2c5120d Merge pull request #277 from iblis17/mac-lapack
add 5d739da ndarray: implement deepcopy via MXNDArrayGetDataNDArray
add f20534d Merge pull request #273 from iblis17/issue-273
add 7cc66d1 executor: fix docstring of `bind`
add a51cba4 Merge pull request #281 from iblis17/bind-doc
add 4e439ed build: remove FileRule on config.mk and fix CUDA config
add d824631 Merge pull request #275 from iblis17/file-rule-cuda
add e3ad29d travis: bring coverage report back
add bbd0e66 Merge pull request #285 from iblis17/cov-report
add 3d4adfb ndarray: `reshape` share the same API with Base
add b68ca2e Merge pull request #272 from iblis17/issue-272
add 16c2c5a base: update GRAD_REQ enumeration
add bb88d79 Merge pull request #283 from iblis17/issue-283
add dcba690 build: CuDNN detection
add 3595403 Merge pull request #280 from iblis17/build-cudnn
add 0b9bf78 sym-node: support `print` for debugging
add df13ddd Merge pull request #276 from iblis17/print-symbol
add 4e89c91 sym-node: extract a until function _create_atomic_symbol
add 63a34cc sym-node: implement reshape api as Base
add 68d6ff7 sym-node: update deprecation for reshape
add a2e270e sym-node: update NEWS for reshape API
add 9e0844a sym-node: add reverse keyword arg
add 55ccbfa sym-node: more test cases for reshape
add 6d1ba53 Merge pull request #279 from iblis17/sym-reshape
add 251a8b3 random: upstream api change
add ebb03d0 Merge pull request #288 from iblis17/upstream-random
add ff0556b random: rename srand! to srand
add 5b3c507 random: update NEWS and deprecated.jl
add 0a53778 Merge pull request #282 from iblis17/srand
add ac7bbd0 base: support MXGetVersion
add 19082ef Merge pull request #289 from iblis17/mx-get-ver
add 904eb20 rand: docstring
add 9d90506 Merge pull request #295 from iblis17/doc-rand
add aae016c ndarray: pretty printing
add 15def2c Merge pull request #290 from iblis17/nd-pretty-print
add 81b134f ndarray: implement `fill` and `fill!` (#297)
add 077a8cb build: improve cuda detection
add a79e33c Merge pull request #287 from iblis17/issue-286
add 71f2d40 base: merge _julia_to_mx_param into dump_mx_param (#296)
add 3824df0 Fix mx.chain var reference in macro (#299)
add 41f8f55 visualize: fix node_attr on upstream master (#301)
add 9fcab40 symbol: add Base.show (#302)
add 4f182ee ndarray: more Base-like APIs (#303)
add a3317f1 ndarray: implement rdiv (#292)
add 8cc5c1c util: docstring of _sig_checker (#305)
add a488d7a ndarray: intro TakingBroadcastSeriously and broadcast unfusing (#300)
add 57cc677 ndarray: elementwise power (#293)
add eea128a ndarray: fix `add` broadcasting and more tests (#308)
add f839be1 Fix deprecations (#307)
add 4919273 ndarray: getindex/setindex! linear indexing (#294)
add 1fc03f2 ndarray: elementwise power for irrational (#310)
add 00d61d2 executor: return `outputs` on forward (#312)
add 3f93ffc git: union merge policy for NEWS.md (#309)
add 678b49e ndarray: regularize fill! API (#311)
add e083e5f doc: cleanup useless enc file (#315)
add 09b9718 build: prevent CUDA_HOME KeyError (#316)
add 9304e6e sym: broadcast unfusion (#314)
add f8e1938 Remove Compat and src/compat.jl (#318)
add a19fc93 test: replace `reldiff` with `isapprox` (#321)
add 526857e docs: address docstring override warning (#320)
add 8071186 travis: enable 'build branch updates' on master/stable only (#319)
add e4b6c0e docs: update examples in overview (#317)
add a51d9b5 docs: fix more example rendering in overview (#322)
add 73b856b Update README badge (#324)
add 1dfb5b3 Cut a release for mxnet v0.12.0 (#304)
add b56c44e Bump upstream to v0.12.1 (#326)
add 9efedbc Update releasing note (#327)
add a4aab01 docs: minor changes in overview (#330)
add 57b50d4 build: set upstream version back to master (#328)
add 8564f19 executor: add Base.show and Base.print (#337)
add ceb7fbf travis: enable caching for reducing build time (#329)
add 1a7887c ndarray: add outer constrcutor for AbstractArray (#334)
add 010ea3c sym: implement get_children (#333)
add 1f50a14 sym: fix printing of symbol generated via get_internals (#332)
add b0556e6 Update regression-example.jl (#339)
add 91a410e test: minor style changes for mlp-test (#340)
add f126482 sym: handy macro `mx.var` for creating `mx.Variable` (#338)
add cb06a9a build: bump max number of make jobs to 32 (#343)
add 45d6279 ndarray: add `Base.show(io, MIME"text/plain")` (#347)
add e0f625a example: fix batchsize config in regression-example (#342)
add 935eb35 ndarray: make _minus type stable (#345)
add 2a5a284 model: refine and test cases for FeedForward constructor (#346)
add cb042fd parametric NDArray (#331)
add 914fab6 ndarray: minor style changes (#349)
add c43d0dd model: fix test cases macro (#355)
add 09ee1f4 ndarray: implement `size(x, dims...)` (#350)
add 8a19651 ndarray: implement endof (#351)
add fd7fb79 ndarray: make _plus type stable (#352)
add b9855c4 ndarray: type convert of _mul_scalar (#356)
add ad57be9 ndarray: type convertion of _plus_scalar (#360)
add 642b17b ndarray: type convertion of _rdiv_scalar (#358)
add ce0e237 doc: sort out table for ndarray api (#362)
add f8d4f62 ndarray: type convertion of _minus_scalar (#354)
add 14f9748 ndarray: do not auto-import _full, _ones and _zeros (#359)
add daf787c ndarray: make trigonometric functions dot-call only (#361)
add 233fcfc ndarray: change internal api of plus to help autograd (#364)
add eb819b0 ndarray: change internal api of mul/div to help autograd (#366)
add 2ca5565 ndarray: change internal api of minus to help autograd (#365)
add 12198f0 ndarray: copy(AbstractArray, context) (#367)
add 2d7cdc6 Reexporting NDArray and SymbolicNode (#306)
add db09528 example: fix parametric NDArray in lstm example (#370)
add 27c66ec sym: overload `bind` (#372)
add bfb1cc4 ndarray: remap hyperbolic function (#374)
add 6609616 ndarray: support transpose on 1D array (#375)
add 881759f example: fix MSE init (#376)
add 5908d97 base: cleanup stale code in _defstruct_impl (#377)
add a941f3a ndarray: add modulo operator (#373)
add 8135a63 ndarray: type convertion of _div_scalar (#357)
add 1e20f50 ndarray: protect from diving zero for non-inplace op (#378)
add d921225 doc/ndarray: add ref table for hyperbolic function (#382)
add 4507598 ndarray: `cat`, `hcat`, `vcat` (#380)
add 39ecb32 base: improve MXError display (#379)
add cb443b7 ndarray: porting Python's autograd (#274)
add 883cdd3 build: make appending lib to MXNET_HOME optional (#386)
add bfeba81 ndarray: support matrix/tensor multiplication (#384)
add 3787895 ndarray: remap expand_dims (#381)
add 813bfdc random: Base-like APIs for rand, rand!, randn, randn! (#383)
add d3cdbf6 ndarray: remap activation functions (#387)
add 9d60663 ndarray: refine signature of `clip` (#391)
add a4143e9 Bump :copyright: (#397)
add ae425bf doc: append 'Defined in' into operator docstring (#393)
add 930090b ndarray: fix 1D NDArray display (#399)
add a1cef7f ndarray: fix type convertion in setindex! (#398)
add 86ffb5d ndarray: inplace modulo operators (#389)
add 30852dd ndarray: support modulo operation in macro inplace (#390)
add 0ef5966 ndarray: add Base.ones(::NDArray) and Base.zeros (#363)
add 49399fc ndarray: broadcasting along dimension on arith operators (#401)
add e79ce71 ndarray: broadcast comparison operators (#402)
add 2ab0584 ndarray: copy!(NDArray, AbstractArray) (#405)
add 1c990bc ndarray: new constructor NDArray(Type, AbstractArray) (#404)
add 0fba766 ndarray: broadcast_min/max (#403)
add 39beaaf remap broadcast_to/axis/axes (#406)
add ec2b699 ndarray: remap hypot (#407)
add 157b088 doc: update predict section of mnist tutorial (#411)
add 9f4f533 Optimizer module overhaul (#396)
add 2c09c26 Update README: `predict` section (#412)
add 6ac37f9 kvstore: refine and copy docstring from Python (#388)
add 4038704 build: Propagate more build flags from ENV (#414)
add 3c8e9d3 build: propagate CC/CXX into config.mk (#419)
add 3df1214 build: propagate USE_JEMALLOC (#422)
add 0ccf8b4 build: add `libmxnet.so` into lib search list. (#424)
add b74a516 Update README.md
add 224ae96 Typo in README
add 06e1d36 Add 'julia/' from commit '224ae962db8ae252ce1ac317c6d4a2cfe67a7d77'
add 9f8425b import Julia binding
add 931555c Merge branch 'pr-10149' import Julia binding
No new revisions were added by this update.
Summary of changes:
Jenkinsfile | 44 +-
ci/docker/Dockerfile.build.ubuntu_cpu | 3 +
ci/docker/install/ubuntu_julia.sh | 37 +
ci/docker/runtime_functions.sh | 54 +
julia/.gitattributes | 1 +
julia/.gitignore | 13 +
julia/.travis.yml | 59 +
julia/LICENSE.md | 179 ++
julia/NEWS.md | 692 ++++++++
julia/README-DEV.md | 13 +
julia/README.md | 60 +
julia/REQUIRE | 7 +
julia/appveyor.yml | 56 +
julia/deps/build.jl | 280 +++
julia/deps/cblas.h | 563 +++++++
julia/deps/cpcblas.sh | 25 +
julia/docs/Makefile | 20 +
julia/docs/make.jl | 29 +
julia/docs/mkdocs.yml | 65 +
julia/docs/src/api.md | 18 +
julia/docs/src/api/callback.md | 6 +
julia/docs/src/api/context.md | 6 +
julia/docs/src/api/executor.md | 6 +
julia/docs/src/api/initializer.md | 6 +
julia/docs/src/api/io.md | 120 ++
julia/docs/src/api/kvstore.md | 6 +
julia/docs/src/api/metric.md | 10 +
julia/docs/src/api/model.md | 9 +
julia/docs/src/api/ndarray.md | 57 +
julia/docs/src/api/nn-factory.md | 9 +
julia/docs/src/api/optimizer.md | 66 +
julia/docs/src/api/symbolic-node.md | 6 +
julia/docs/src/api/visualize.md | 6 +
julia/docs/src/index.md | 55 +
julia/docs/src/tutorial/char-lstm.md | 306 ++++
julia/docs/src/tutorial/images/LSTM3-chain.png | Bin 0 -> 229688 bytes
julia/docs/src/tutorial/images/char-lstm-vis.svg | 435 +++++
julia/docs/src/tutorial/mnist.md | 264 +++
julia/docs/src/user-guide/faq.md | 8 +
julia/docs/src/user-guide/install.md | 92 +
julia/docs/src/user-guide/overview.md | 406 +++++
julia/examples/char-lstm/.gitignore | 6 +
julia/examples/char-lstm/README.md | 121 ++
julia/examples/char-lstm/config.jl | 40 +
julia/examples/char-lstm/lstm.jl | 166 ++
julia/examples/char-lstm/sampler.jl | 89 +
julia/examples/char-lstm/seq-data.jl | 136 ++
julia/examples/char-lstm/train.jl | 59 +
julia/examples/char-lstm/visualize.jl | 32 +
julia/examples/cifar10/cifar10.jl | 101 ++
.../Prediction with Pre-trained Model.ipynb | 249 +++
.../imagenet/ijulia-pretrained-predict/cat.png | Bin 0 -> 123126 bytes
.../ijulia-pretrained-predict/imagehelper.py | 48 +
julia/examples/mnist/lenet-stn.jl | 81 +
julia/examples/mnist/lenet.jl | 62 +
julia/examples/mnist/mlp-test.jl | 122 ++
julia/examples/mnist/mlp.jl | 75 +
julia/examples/mnist/mnist-data.jl | 33 +
julia/examples/nondefault-example.jl | 147 ++
julia/examples/regression-example.jl | 101 ++
julia/models/Inception/.gitignore | 2 +
julia/models/Inception/get.sh | 22 +
julia/plugins/README.md | 14 +
julia/plugins/io/svmlight.jl | 86 +
julia/src/MXNet.jl | 173 ++
julia/src/autograd.jl | 404 +++++
julia/src/base.jl | 317 ++++
julia/src/broadcast.jl | 36 +
julia/src/callback.jl | 160 ++
julia/src/context.jl | 54 +
julia/src/deprecated.jl | 160 ++
julia/src/executor.jl | 254 +++
julia/src/initializer.jl | 197 +++
julia/src/io.jl | 643 +++++++
julia/src/kvstore.jl | 353 ++++
julia/src/metric.jl | 478 ++++++
julia/src/model.jl | 671 ++++++++
julia/src/name.jl | 61 +
julia/src/ndarray.jl | 1780 ++++++++++++++++++++
julia/src/nn-factory.jl | 60 +
julia/src/optimizer.jl | 308 ++++
julia/src/optimizers/adadelta.jl | 104 ++
julia/src/optimizers/adagrad.jl | 77 +
julia/src/optimizers/adam.jl | 88 +
julia/src/optimizers/adamax.jl | 84 +
julia/src/optimizers/nadam.jl | 111 ++
julia/src/optimizers/rmsprop.jl | 84 +
julia/src/optimizers/sgd.jl | 88 +
julia/src/random.jl | 83 +
julia/src/symbolic-node.jl | 1003 +++++++++++
julia/src/util.jl | 255 +++
julia/src/visualize.jl | 213 +++
julia/test/common.jl | 50 +
julia/test/runtests.jl | 43 +
julia/test/travis/run_coverage.sh | 18 +
julia/test/travis/run_test.sh | 28 +
julia/test/travis/setup_env.sh | 38 +
julia/test/unittest/autograd.jl | 403 +++++
julia/test/unittest/bind.jl | 95 ++
julia/test/unittest/initializer.jl | 35 +
julia/test/unittest/io.jl | 145 ++
julia/test/unittest/kvstore.jl | 114 ++
julia/test/unittest/metric.jl | 116 ++
julia/test/unittest/model.jl | 51 +
julia/test/unittest/name.jl | 50 +
julia/test/unittest/ndarray.jl | 1475 ++++++++++++++++
julia/test/unittest/operator.jl | 57 +
julia/test/unittest/optimizer.jl | 85 +
julia/test/unittest/random.jl | 62 +
julia/test/unittest/symbolic-node.jl | 565 +++++++
julia/test/unittest/util.jl | 51 +
julia/test/unittest/visualize.jl | 54 +
.../nightly/apache_rat_license_check/rat-excludes | 1 +
tools/license_header.py | 6 +-
114 files changed, 17719 insertions(+), 11 deletions(-)
create mode 100755 ci/docker/install/ubuntu_julia.sh
create mode 100644 julia/.gitattributes
create mode 100644 julia/.gitignore
create mode 100644 julia/.travis.yml
create mode 100644 julia/LICENSE.md
create mode 100644 julia/NEWS.md
create mode 100644 julia/README-DEV.md
create mode 100644 julia/README.md
create mode 100644 julia/REQUIRE
create mode 100644 julia/appveyor.yml
create mode 100644 julia/deps/build.jl
create mode 100644 julia/deps/cblas.h
create mode 100755 julia/deps/cpcblas.sh
create mode 100644 julia/docs/Makefile
create mode 100644 julia/docs/make.jl
create mode 100644 julia/docs/mkdocs.yml
create mode 100644 julia/docs/src/api.md
create mode 100644 julia/docs/src/api/callback.md
create mode 100644 julia/docs/src/api/context.md
create mode 100644 julia/docs/src/api/executor.md
create mode 100644 julia/docs/src/api/initializer.md
create mode 100644 julia/docs/src/api/io.md
create mode 100644 julia/docs/src/api/kvstore.md
create mode 100644 julia/docs/src/api/metric.md
create mode 100644 julia/docs/src/api/model.md
create mode 100644 julia/docs/src/api/ndarray.md
create mode 100644 julia/docs/src/api/nn-factory.md
create mode 100644 julia/docs/src/api/optimizer.md
create mode 100644 julia/docs/src/api/symbolic-node.md
create mode 100644 julia/docs/src/api/visualize.md
create mode 100644 julia/docs/src/index.md
create mode 100644 julia/docs/src/tutorial/char-lstm.md
create mode 100644 julia/docs/src/tutorial/images/LSTM3-chain.png
create mode 100644 julia/docs/src/tutorial/images/char-lstm-vis.svg
create mode 100644 julia/docs/src/tutorial/mnist.md
create mode 100644 julia/docs/src/user-guide/faq.md
create mode 100644 julia/docs/src/user-guide/install.md
create mode 100644 julia/docs/src/user-guide/overview.md
create mode 100644 julia/examples/char-lstm/.gitignore
create mode 100644 julia/examples/char-lstm/README.md
create mode 100644 julia/examples/char-lstm/config.jl
create mode 100644 julia/examples/char-lstm/lstm.jl
create mode 100644 julia/examples/char-lstm/sampler.jl
create mode 100644 julia/examples/char-lstm/seq-data.jl
create mode 100644 julia/examples/char-lstm/train.jl
create mode 100644 julia/examples/char-lstm/visualize.jl
create mode 100644 julia/examples/cifar10/cifar10.jl
create mode 100644 julia/examples/imagenet/ijulia-pretrained-predict/Prediction with Pre-trained Model.ipynb
create mode 100644 julia/examples/imagenet/ijulia-pretrained-predict/cat.png
create mode 100644 julia/examples/imagenet/ijulia-pretrained-predict/imagehelper.py
create mode 100644 julia/examples/mnist/lenet-stn.jl
create mode 100644 julia/examples/mnist/lenet.jl
create mode 100644 julia/examples/mnist/mlp-test.jl
create mode 100644 julia/examples/mnist/mlp.jl
create mode 100644 julia/examples/mnist/mnist-data.jl
create mode 100644 julia/examples/nondefault-example.jl
create mode 100644 julia/examples/regression-example.jl
create mode 100644 julia/models/Inception/.gitignore
create mode 100755 julia/models/Inception/get.sh
create mode 100644 julia/plugins/README.md
create mode 100644 julia/plugins/io/svmlight.jl
create mode 100644 julia/src/MXNet.jl
create mode 100644 julia/src/autograd.jl
create mode 100644 julia/src/base.jl
create mode 100644 julia/src/broadcast.jl
create mode 100644 julia/src/callback.jl
create mode 100644 julia/src/context.jl
create mode 100644 julia/src/deprecated.jl
create mode 100644 julia/src/executor.jl
create mode 100644 julia/src/initializer.jl
create mode 100644 julia/src/io.jl
create mode 100644 julia/src/kvstore.jl
create mode 100644 julia/src/metric.jl
create mode 100644 julia/src/model.jl
create mode 100644 julia/src/name.jl
create mode 100644 julia/src/ndarray.jl
create mode 100644 julia/src/nn-factory.jl
create mode 100644 julia/src/optimizer.jl
create mode 100644 julia/src/optimizers/adadelta.jl
create mode 100644 julia/src/optimizers/adagrad.jl
create mode 100644 julia/src/optimizers/adam.jl
create mode 100644 julia/src/optimizers/adamax.jl
create mode 100644 julia/src/optimizers/nadam.jl
create mode 100644 julia/src/optimizers/rmsprop.jl
create mode 100644 julia/src/optimizers/sgd.jl
create mode 100644 julia/src/random.jl
create mode 100644 julia/src/symbolic-node.jl
create mode 100644 julia/src/util.jl
create mode 100644 julia/src/visualize.jl
create mode 100644 julia/test/common.jl
create mode 100644 julia/test/runtests.jl
create mode 100644 julia/test/travis/run_coverage.sh
create mode 100755 julia/test/travis/run_test.sh
create mode 100755 julia/test/travis/setup_env.sh
create mode 100644 julia/test/unittest/autograd.jl
create mode 100644 julia/test/unittest/bind.jl
create mode 100644 julia/test/unittest/initializer.jl
create mode 100644 julia/test/unittest/io.jl
create mode 100644 julia/test/unittest/kvstore.jl
create mode 100644 julia/test/unittest/metric.jl
create mode 100644 julia/test/unittest/model.jl
create mode 100644 julia/test/unittest/name.jl
create mode 100644 julia/test/unittest/ndarray.jl
create mode 100644 julia/test/unittest/operator.jl
create mode 100644 julia/test/unittest/optimizer.jl
create mode 100644 julia/test/unittest/random.jl
create mode 100644 julia/test/unittest/symbolic-node.jl
create mode 100644 julia/test/unittest/util.jl
create mode 100644 julia/test/unittest/visualize.jl