From a21f6286a91d718e8f6889db96c09da63d860f1a Mon Sep 17 00:00:00 2001 From: zhangli Date: Tue, 14 Mar 2017 16:37:35 +0800 Subject: [PATCH 001/314] refactor From fa197ab1325de564503f9606c6f89e271f328111 Mon Sep 17 00:00:00 2001 From: zhangli Date: Tue, 21 Mar 2017 09:25:06 +0800 Subject: [PATCH 002/314] merge visualization From 926f5b99d41d879544122ff7a3cb62f452954a72 Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Wed, 22 Mar 2017 10:04:22 +0800 Subject: [PATCH 003/314] add Merge Module to the constructor to support multi-layer BiRNN From 46d8bceccdf39a35230c003444d3c3635c060c6d Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Wed, 22 Mar 2017 12:53:20 +0800 Subject: [PATCH 004/314] add a default null for constuctor From 10d960b98db28544455403f9f3d4a496600a12bf Mon Sep 17 00:00:00 2001 From: zhangli Date: Tue, 7 Mar 2017 10:33:41 +0800 Subject: [PATCH 005/314] add module.predict From 66fd888294f0ca690218de76cc9a12d2c0a2f654 Mon Sep 17 00:00:00 2001 From: shenyurun <491507534@qq.com> Date: Thu, 9 Mar 2017 11:17:06 +0800 Subject: [PATCH 006/314] Add sizeAverage for TimeDistributedCriterion Remove unnecessary variables and methods update timedistributed Update TimeDistributedCriterion Update license Remove if() in extend From 8d700bbf15c6252d809aa56630bbf38ea9f5b741 Mon Sep 17 00:00:00 2001 From: Yao Date: Fri, 17 Feb 2017 02:42:08 +0800 Subject: [PATCH 007/314] add buildHidden to LSTM add bulidCell to the LSTM add buildLSTM finish lstm Made some modification on Reccurent to make it work for LSTM and other table layers delete some redundant code make lstm work add gradient check for lstm add parameter methods to LSTM and add more tests to NarrowTable and SelectTable make LSTM work correctly add element-research rnn comparison test fix reccurent forward bug and set timeDistributedCriterion sizeAverage to false, passed the test with Element-Research add SplitTable make lstm work add more tests for lstm Merge with recurrent unchange sequential layer add comments to lstm remove unnecessary change meet the code review add parallel tag to LSTMPeepholeSpec remove dropout arguments make batch lstm pass tests make LSTMPeephole pass the test From de51a6c2c7fd35865bfbc0d260bdc0bb8a9d74c8 Mon Sep 17 00:00:00 2001 From: Yao Date: Fri, 24 Mar 2017 10:09:16 +0800 Subject: [PATCH 008/314] add LSTMPeephole python support fix conflicts remove core add core back fix style check fix random failure From e4ef14aec01894592392d97b345c2ef5b6364b2c Mon Sep 17 00:00:00 2001 From: zhangli Date: Wed, 22 Mar 2017 15:32:20 +0800 Subject: [PATCH 009/314] add scala doc and python doc From 3ea529007e9811d1a4a8a064006596adbe729276 Mon Sep 17 00:00:00 2001 From: qiuxin2012 Date: Tue, 28 Mar 2017 11:07:15 +0800 Subject: [PATCH 010/314] delete copy From bd8ff151d409d76d5651803ce0019446cf4d7958 Mon Sep 17 00:00:00 2001 From: zhangli Date: Fri, 3 Mar 2017 14:33:55 +0800 Subject: [PATCH 011/314] add Validator test From 1279ddd779f0a2cf83852cc2b6f3d33d55e37bd3 Mon Sep 17 00:00:00 2001 From: qiuxin2012 Date: Tue, 28 Mar 2017 15:48:58 +0800 Subject: [PATCH 012/314] add some scala doc From 654e1b979d13f36302fc13cb2ea101666c4ea5a3 Mon Sep 17 00:00:00 2001 From: Yao Date: Tue, 28 Mar 2017 09:07:06 +0800 Subject: [PATCH 013/314] add scala documentation From 94f513dde5042f59f19eb3cbab57ef3e61dbf84e Mon Sep 17 00:00:00 2001 From: Yao Date: Tue, 28 Mar 2017 09:19:07 +0800 Subject: [PATCH 014/314] add some more scala doc From 25c52f03c9983ee8cc5130a0ffeb69c1301c6f10 Mon Sep 17 00:00:00 2001 From: Yao Date: Wed, 29 Mar 2017 03:47:22 +0800 Subject: [PATCH 015/314] meet code review From fa0de39ebc5ca2209009e364a8dc5f6839887ee9 Mon Sep 17 00:00:00 2001 From: zhangli Date: Tue, 28 Mar 2017 14:42:52 +0800 Subject: [PATCH 016/314] add some docs From 70a57a22e6e02f70f3b7738a09d70c2ca77f21ab Mon Sep 17 00:00:00 2001 From: plrectco Date: Wed, 22 Mar 2017 10:10:32 +0800 Subject: [PATCH 017/314] add python and scala docs From 267cdfecb56fa9a7061dbd940858bfcabda7d0b0 Mon Sep 17 00:00:00 2001 From: seaofocean Date: Wed, 29 Mar 2017 12:17:07 +0800 Subject: [PATCH 018/314] add scala doc From dc35750ee3feec144ed7fc28f5febd5d13de108f Mon Sep 17 00:00:00 2001 From: ian Date: Wed, 29 Mar 2017 21:27:04 +0800 Subject: [PATCH 019/314] add doc From ba706ab6bfe72c4658762da3a9019507c67c4568 Mon Sep 17 00:00:00 2001 From: ian Date: Thu, 30 Mar 2017 12:55:07 +0800 Subject: [PATCH 020/314] meet code review From afffd380f405015bf490acf212a57ea4322a1170 Mon Sep 17 00:00:00 2001 From: Yao Date: Wed, 29 Mar 2017 03:21:26 +0800 Subject: [PATCH 021/314] add some python doc From 4e5a4eed70383e0df8e199229310c3b321dd4d06 Mon Sep 17 00:00:00 2001 From: plrectco Date: Thu, 30 Mar 2017 14:09:41 +0800 Subject: [PATCH 022/314] add softmargincriterion change spelling of rnn From 8af37fb3d1641b8ac097cac37340bf5a514a5411 Mon Sep 17 00:00:00 2001 From: plrectco Date: Thu, 30 Mar 2017 13:44:01 +0800 Subject: [PATCH 023/314] update mixturetable From b5667da38380ba46c647371e21283ea64dbc1dfd Mon Sep 17 00:00:00 2001 From: Yao Date: Fri, 31 Mar 2017 07:42:09 +0800 Subject: [PATCH 024/314] add more comments add more python doc fix scala style From 2cb8c4733dba3901030db47f0e36fa65313e2c59 Mon Sep 17 00:00:00 2001 From: qiuxin2012 Date: Fri, 31 Mar 2017 10:45:03 +0800 Subject: [PATCH 025/314] add python doc From 39d200138e348d04b6846e2aed36f71e9fd5cee8 Mon Sep 17 00:00:00 2001 From: shenyurun <491507534@qq.com> Date: Thu, 30 Mar 2017 15:22:19 +0800 Subject: [PATCH 026/314] Add some docs From 83fe17e56dbccf9f7c2e0ef241cf0445bf59a86b Mon Sep 17 00:00:00 2001 From: shenyurun <491507534@qq.com> Date: Fri, 31 Mar 2017 15:49:44 +0800 Subject: [PATCH 027/314] modify more docs and remove unnecessary parameters From bee2c6785b4f39fd2a1e0f0ea9ea54154e9100bc Mon Sep 17 00:00:00 2001 From: Yao Date: Wed, 22 Feb 2017 09:52:29 +0800 Subject: [PATCH 028/314] add accGradParameters to Sequential and test it From 48f790691602e70bc4e1825e9eca405d4a0b67e2 Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Sat, 1 Apr 2017 10:53:09 +0800 Subject: [PATCH 029/314] clear the weights then clone to save memory From a490064b848157cd58a9e01e2c2c2b1114268b4f Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Sat, 1 Apr 2017 10:56:14 +0800 Subject: [PATCH 030/314] clear the gradWeights From 7da379d25f8543f4b09c44d756404bc1f0302669 Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Sat, 1 Apr 2017 15:03:46 +0800 Subject: [PATCH 031/314] cloned module weights.set() From 54db85371764912c37d1ad584ebb13ba9de5ecd1 Mon Sep 17 00:00:00 2001 From: Yao Date: Sat, 1 Apr 2017 08:46:58 +0800 Subject: [PATCH 032/314] add backward to sequential refine the test From a32e0297508091db5c4cf23f4fd0db8a2637c5ca Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Fri, 7 Apr 2017 11:03:27 +0800 Subject: [PATCH 033/314] remove Array[Future]; make it run in single thread From 80db4b9ab578a110594c682b031d443766e162d2 Mon Sep 17 00:00:00 2001 From: Yao Date: Wed, 22 Mar 2017 07:08:33 +0800 Subject: [PATCH 034/314] Make Recurrent support lazy Dropout From efb7b097cf90bc1ab35648de9459bec101eaf2e6 Mon Sep 17 00:00:00 2001 From: Yao Date: Wed, 22 Mar 2017 09:09:02 +0800 Subject: [PATCH 035/314] Add Dropout api to LSTM fig a bug in PythonBigDL fix a bug in LSTMSpec From db5a34b7805cab2c52ebaef6635068eadbd33b05 Mon Sep 17 00:00:00 2001 From: Yao Date: Thu, 23 Mar 2017 02:42:15 +0800 Subject: [PATCH 036/314] add comments to Dropout From 8efdb5c653f86c812f384fb5997b46d47b4aaca5 Mon Sep 17 00:00:00 2001 From: Yao Date: Thu, 23 Mar 2017 06:29:28 +0800 Subject: [PATCH 037/314] Add test of recurrent dropout From c6de79c903529b41729136bd93d49bf482ef2866 Mon Sep 17 00:00:00 2001 From: Yao Date: Thu, 23 Mar 2017 10:11:02 +0800 Subject: [PATCH 038/314] fix a bug for droput when input is not contiguous From 6e6262d58c3f2624d144485a225680cf00770aa2 Mon Sep 17 00:00:00 2001 From: Yao Date: Fri, 24 Mar 2017 07:08:14 +0800 Subject: [PATCH 039/314] remove isLazy and flag from Dropout and add isResampleing meet review From d6215bff3d28d14a2bd7e4623413cc3fe22751f3 Mon Sep 17 00:00:00 2001 From: Yao Date: Fri, 24 Mar 2017 09:49:13 +0800 Subject: [PATCH 040/314] add more docuemnts to Cell class From d14969f7d835f2a3dae82765dddd18ed191af54c Mon Sep 17 00:00:00 2001 From: Yao Date: Thu, 6 Apr 2017 09:41:28 +0800 Subject: [PATCH 041/314] add dropout to lstmPeepwhole From 8fa9229208c59153f04bff86c121a5459cd4bbf4 Mon Sep 17 00:00:00 2001 From: Yao Date: Tue, 11 Apr 2017 03:08:24 +0800 Subject: [PATCH 042/314] add more comments From 1ecfdad0c2111866358431db5c64320790f091ef Mon Sep 17 00:00:00 2001 From: seaofocean Date: Wed, 12 Apr 2017 18:57:09 +0800 Subject: [PATCH 043/314] add VolumetricConvolution From ee6adf0813d5cf59996f60704ce5a1d795d876e0 Mon Sep 17 00:00:00 2001 From: Yiheng Wang Date: Tue, 11 Apr 2017 13:41:12 +0800 Subject: [PATCH 044/314] checkin graph From 790c2569c6e488e763c9f86dcce5cc7b3f1d048a Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Tue, 18 Apr 2017 14:55:43 +0800 Subject: [PATCH 045/314] support DiceCoefficient From c225959f285316d3bac4a89dffe465db2ab75ae3 Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Tue, 18 Apr 2017 16:58:59 +0800 Subject: [PATCH 046/314] revise Parallel to serial in spec gradient check for DiceCoefficient From 069c023a5054d2db562c706d3f82f65b2d2aeb68 Mon Sep 17 00:00:00 2001 From: zhangli Date: Wed, 5 Apr 2017 13:17:31 +0800 Subject: [PATCH 047/314] add evaluator From 3c3bae9361256f643a10ab7ad8b2aef785736a4d Mon Sep 17 00:00:00 2001 From: plrectco Date: Wed, 19 Apr 2017 16:05:50 +0800 Subject: [PATCH 048/314] change some vars to vals From 2d917167fda3e6d199c28cbfe5197a1475fcbaa2 Mon Sep 17 00:00:00 2001 From: seaofocean Date: Tue, 25 Apr 2017 12:46:00 +0800 Subject: [PATCH 049/314] assign updateOutput return value to output From 3e95bab852a212e14686dabc4503b1ca7ba48a50 Mon Sep 17 00:00:00 2001 From: zhangli Date: Tue, 11 Apr 2017 09:37:28 +0800 Subject: [PATCH 050/314] add docs for lookuptable From 080d726bd884cabc5a03b14611d2f62db1c08132 Mon Sep 17 00:00:00 2001 From: seaofocean Date: Tue, 18 Apr 2017 12:18:52 +0800 Subject: [PATCH 051/314] remove unnecessary clone in FlattenTable From 57bcf133005f152393517725c08385e055d33cf8 Mon Sep 17 00:00:00 2001 From: seaofocean Date: Mon, 17 Apr 2017 15:43:36 +0800 Subject: [PATCH 052/314] add 4d normalize support From 27b8319723d198c26cca9644bdab768d1cc3a07f Mon Sep 17 00:00:00 2001 From: Yiheng Wang Date: Wed, 26 Apr 2017 12:59:38 +0800 Subject: [PATCH 053/314] refactor squeeze layer From bd61037ded0dad99a33c33bc048a204a8e728893 Mon Sep 17 00:00:00 2001 From: yangw Date: Tue, 25 Apr 2017 11:27:56 +0800 Subject: [PATCH 054/314] update biRecurrent and add adam optim method in python api From 442ea611122858a418b13a9fe6bbead9c6527ccf Mon Sep 17 00:00:00 2001 From: Xianyan Date: Fri, 28 Apr 2017 17:30:42 +0800 Subject: [PATCH 055/314] add VolumetricMaxPooling (#829) * add VolumetricMaxPooling From 75c96b523dbccaeeccb41f90a1e98bd6f9c84471 Mon Sep 17 00:00:00 2001 From: Yiheng Wang Date: Thu, 4 May 2017 17:33:22 +0800 Subject: [PATCH 056/314] revise Graph and DirectedGraph From 708cbe73f29e895e3a6066b855125a0ad913078e Mon Sep 17 00:00:00 2001 From: jenniew Date: Thu, 6 Apr 2017 15:49:11 -0700 Subject: [PATCH 057/314] set weights per layer remove return value of set weights update setWeights in scala side and add negative test case update setweights From 6548dcd66e9f59f541310101a3f0aa911c0312ae Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Mon, 8 May 2017 16:26:52 +0800 Subject: [PATCH 058/314] Support layerwise, weight-bias separated and multiple sorts (L1, L2 and others) of regularization (#714) * make linear layerwise regurization work * add regularizer option to layers with weights fix a bug * new version * include accRegularization in accGradparameters * meet code review * add layerwise regualrization performance tests * fix to meet code review * make Regularizer.isRegularized private and optimize L1 regularization implementation * fix jenkins tests failed * change l1 implementation * improve layerwise regularization performance test * improve layerwise regularization performance test * delete layerwise regularization perform test From 6c7506541ebaeec049f30ace058553780fcae555 Mon Sep 17 00:00:00 2001 From: jenniew Date: Fri, 5 May 2017 15:12:39 -0700 Subject: [PATCH 059/314] update style and set/get weights method name remove not used import From 16eafb4e5200b5ae22c0b7c1fa1779ddc9511afa Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 11 May 2017 14:40:27 +0800 Subject: [PATCH 060/314] fix bug of CMul and CAdd when getParameters is called (#872) From 446e774c012ac88a5a215b4b906609f39e52674f Mon Sep 17 00:00:00 2001 From: SujieZhu Date: Thu, 18 May 2017 15:20:24 +0800 Subject: [PATCH 061/314] fix the Squeeze layer for multi dimension From ba1f202d8b51e053b4e1409c61673bce205e78d4 Mon Sep 17 00:00:00 2001 From: seaofocean Date: Thu, 18 May 2017 17:20:30 +0800 Subject: [PATCH 062/314] fix the bug of SpatialDilatedConvolution From e1d987f458c85b2a5741387fab70cfc86c51c3dc Mon Sep 17 00:00:00 2001 From: "li,zhichao" Date: Fri, 19 May 2017 14:23:19 +0800 Subject: [PATCH 063/314] Refine model name (#912) * refine model name * add printName * style From 52cc3d2fbee427f95ac55728ff3e2574460026bf Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Mon, 8 May 2017 16:53:50 +0800 Subject: [PATCH 064/314] linear; initialize gradWeight when running backward From 00d473811a5d3634f4045cabbb2e9a716aa2b0f1 Mon Sep 17 00:00:00 2001 From: qiuxin2012 Date: Wed, 24 May 2017 14:52:59 +0800 Subject: [PATCH 065/314] fix ConcatTable.clearState() don't clear gradInput when gradInput is a table From fc15ce452183626bbe32fcb6946e6ddad52e9e3b Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Thu, 25 May 2017 10:48:32 +0800 Subject: [PATCH 066/314] fix bug in Concat layer and Add backward to ConcatTable (#902) From 9c39513289458b0b7589c0b7c9e186f43213e14c Mon Sep 17 00:00:00 2001 From: jenniew Date: Wed, 31 May 2017 15:37:49 -0700 Subject: [PATCH 067/314] add tensor size info when requirement failed (#942) * add tensor size info when requirement failed * add layer info when requirement fail From 776009b736a552732ead9ee767607e5beb9cebef Mon Sep 17 00:00:00 2001 From: qiuxin2012 Date: Mon, 5 Jun 2017 13:26:19 +0800 Subject: [PATCH 068/314] fix BCECriterion From 7b9e5cbd15aad40a49945c1f586c4262fc930d64 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Tue, 6 Jun 2017 14:52:08 +0800 Subject: [PATCH 069/314] share memory in CAddTable for LSTM, GRU, SimpleRNN (#855) * share memory in CAddTable for LSTM, GRU, SimpleRNN; share memory for Reverse layer * revise javadoc * revert GRU * set default inPlace value to be false in Reverse layer From 65afa653bef9f1e6a26ef4debc11445f8700038b Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Tue, 6 Jun 2017 15:34:20 +0800 Subject: [PATCH 070/314] fix add bug (#968) From 8437b7b12c04bbe1539dbfacfae45ec540f67667 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Wed, 7 Jun 2017 09:44:31 +0800 Subject: [PATCH 071/314] change SelectTable from Container to AbstractModule (#977) From 1342865e4dfe8db3fe9e0689fa35af3bb64411fb Mon Sep 17 00:00:00 2001 From: Fu Zhouwang Date: Wed, 7 Jun 2017 15:08:20 +0800 Subject: [PATCH 072/314] Add isTensor and isTable method in Activity with additional test cases (#980) * Add isTensor and isTable method in Activity * fix sytle in comment From f6afb611b0919204e1280c09a7d44c07b7874fba Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Wed, 7 Jun 2017 18:08:12 +0800 Subject: [PATCH 073/314] refactor InitializationMethod (#960) * refactor init * fix test * address some comments * fix tests * add trait * fix tests From 17349e48d9ce2ddbb1e9076d7b8e8cc1027762f3 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 7 Jun 2017 19:10:34 +0800 Subject: [PATCH 074/314] some new layer and existing layer refactor (#957) * nn refactor * fix code style issue * add python api * change back the layers From 970763f895da240f0883cbad1ded9860f38e0ea1 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Thu, 8 Jun 2017 09:51:02 +0800 Subject: [PATCH 075/314] delete useless CriterionTable (#981) From c58a6ab4201ffaa018ee150dbe35eb23dccca5cd Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 8 Jun 2017 12:36:15 +0800 Subject: [PATCH 076/314] refactor smoothL1Criterion and fix tensor storage usage (#986) * refactor smoothL1Criterion and fix tensor storage usage From 54c1021b0d50c0827b2bd87032730a66a97d856f Mon Sep 17 00:00:00 2001 From: yangw Date: Thu, 8 Jun 2017 15:05:48 +0800 Subject: [PATCH 077/314] add python api From c67f7f4a76bab76a9240d6cd3670faff2ff7fd88 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Fri, 9 Jun 2017 16:37:50 +0800 Subject: [PATCH 078/314] unchange Euclidean input (#991) From 98535688fe0898e70d1e80dfcee17cdd1c0ce60e Mon Sep 17 00:00:00 2001 From: qiuxin2012 Date: Wed, 14 Jun 2017 19:27:52 +0800 Subject: [PATCH 079/314] fix concat return incorrect result when output is 1D From 5cf02b6bc0f205067960ec9ad91119a1682d3f2c Mon Sep 17 00:00:00 2001 From: "li,zhichao" Date: Fri, 16 Jun 2017 11:26:39 +0800 Subject: [PATCH 080/314] Fix HardShrink factory method (#1025) * fix HardShrink factory * update unittest From 9e94774514816d681383f044f5d8b0cbd2215992 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Tue, 20 Jun 2017 14:41:51 +0800 Subject: [PATCH 081/314] using set instead of copy for output and gradInput (#1028) * using set instead of copy for output and gradInput define private method * revise set method * add condition before set method * grammar revision * set cellAppendStartIdx to cells.length * GRU, last layer CAddTable(false) From 016b66817196002425f8a82a9b56c08f056b26ac Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 21 Jun 2017 12:50:17 +0800 Subject: [PATCH 082/314] Support Tensorflow model file read/write (#800) * nn refactor * nn refactor * fix code style issue * change back the layers * code refactor * change tests to automatic test * add more test for model save * remove some useless unit test * add more save test * add more writer test * rnn test case automation * refine save test * refine save unit test * remove NHWC * meet code review * meet code review * use MulConst in MulTF * add a flatten node for tf 1.1 * fix code style and failed unit test * mv tf model layers to another package * add a python example to show how to define model in tensorflow and run in BigDL * move tf example python code to example folder * Add backward test in LeNet and AlexNet (#19) * add unit test of lenet backward * add some print * add backward test in lenet and alexnet * seperate testModel into forward and backward methods From bf25d2de729dce3e2ad5c22f0e929f13b22cf415 Mon Sep 17 00:00:00 2001 From: Nisha Muktewar Date: Tue, 20 Jun 2017 22:21:24 -0700 Subject: [PATCH 083/314] Patch to improve error message when the target (label) value starts at 0 (#1049) * Changes to SpatialConvolution.scala to avoid divide by 0 error * reversing changes * Update error mesg when target value starts at 0 From 1525c950d0afd6064f92ddf5c2f41e52f1207cc3 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Wed, 21 Jun 2017 13:22:12 +0800 Subject: [PATCH 084/314] update some doc (#1050) From f695785ab0ced9a5a9dd35f52f096f745df97352 Mon Sep 17 00:00:00 2001 From: Yuchao-Tao Date: Thu, 22 Jun 2017 09:24:02 +0800 Subject: [PATCH 085/314] Python regularizer (#1008) * # This is a combination of 5 commits. # The first commit's message is: add Regularizer to PythonBidgDL.scala and optimizer.py # This is the 2nd commit message: update regularizer in layer.py # This is the 3rd commit message: some fix # This is the 4th commit message: some fix 2 # This is the 5th commit message: adjust according to the test * Resolve Conflicts; fix; fix RNN; fix; pass compilation; fix; fix on the missmatch on the parameter list * fix style * fix style * fix style * fix style From 0730824d61b71b186cf624698d18ec2208db64bf Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 22 Jun 2017 11:50:07 +0800 Subject: [PATCH 086/314] [Issue 1035] Allow layer regulazier set after model definition, add regulazier for cadd and cmul (#1048) * Allow layer regulazier set after model definition * add regulazier support for cadd and cmul * meet code review * add python set API, add regularizer to python cmul and cadd * add test for setBRegularizer * add parameters to create method, add doc From 9a643ab0c8bd51698986af7e13015dd17ee88564 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Thu, 22 Jun 2017 16:33:53 +0800 Subject: [PATCH 087/314] change Table to Tensor in L1HingeEmbeddingCriterion (#1061) From 1210ae50feaf9ed170494fe6c4204a34af4fe351 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Fri, 23 Jun 2017 10:10:57 +0800 Subject: [PATCH 088/314] Caffe Model Support - Load Caffe Dynamically from Prototxt and Binary (#964) * add dynamic caffe loading support * refined with bug fix and more commnets * fix typo and bias check for convolution * renaming change * fix import reference * refine customized layer support * fix style check issue * resolve style issue From ff2ada5f8726a50092a5ca5ff929cbc691c448ad Mon Sep 17 00:00:00 2001 From: Xianyan Date: Fri, 23 Jun 2017 10:17:16 +0800 Subject: [PATCH 089/314] let getNamedModule also return containers, add unit test for it (#1067) From ca2dce6049558588e0113f3f037b835046f32127 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Fri, 23 Jun 2017 13:40:19 +0800 Subject: [PATCH 090/314] rename apply to inputs when define connection in Graph API (#1069) * rename apply to inputs * fix style check From df08717f074644e1b98755590b40a0f28c7bcc43 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Fri, 23 Jun 2017 16:52:47 +0800 Subject: [PATCH 091/314] add L1HingeEmbeddingCriterion unit test for python (#1073) From 20517a36c74f14f5874dcb2a70f098c44f7a4656 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Mon, 26 Jun 2017 14:07:58 +0800 Subject: [PATCH 092/314] Polish InferReshape documentation (#1087) * Polish InferReshape documentation * remove linearSpec unnecessary change * fix core conflict * meet code review From ae0ba8a35de901b4455e15446ab918654f815677 Mon Sep 17 00:00:00 2001 From: xinyue Date: Tue, 27 Jun 2017 00:13:10 -0500 Subject: [PATCH 093/314] [Issue 930] Separate scale into scaleW and scaleB (#975) From 96528b334f793d1dec17909a15845d0f80a3b0f9 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 29 Jun 2017 16:02:47 +0800 Subject: [PATCH 094/314] add preTopology to boost performance of rnn (#1117) * add preTopology to boost performance of rnn * require only one Cell added * require logic revision, modules.length == 1 || modules.length == 2 * add Javadoc for preTopology in Cell and Recurrent From a75550b1da40737df7f39294da4946b464d4024a Mon Sep 17 00:00:00 2001 From: yangw Date: Wed, 28 Jun 2017 12:54:27 +0800 Subject: [PATCH 095/314] fix RnnCell factory method From 0fea118680b064ad65e50904fdce2aa06387333b Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Fri, 30 Jun 2017 13:37:50 +0800 Subject: [PATCH 096/314] makedown docs for some API (#1075) * add some api md * change according to reviews From 6c09fdc6445faa2015200be7d220fa9e26484222 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Fri, 30 Jun 2017 13:46:11 +0800 Subject: [PATCH 097/314] fix some unit test (#1104) From 9d9ea0c2f4dbaa4688fc331579a695632bfaed5a Mon Sep 17 00:00:00 2001 From: Xianyan Date: Fri, 30 Jun 2017 15:41:47 +0800 Subject: [PATCH 098/314] fix VolumetricConvolution unit test, and fix bug in batch mode (#1105) * fix VolumetricConvolution unit test, and fix bug in batch mode * meet pr review From 44d65c765b3b1ea8c56982cdf54fd67c02bcee02 Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Mon, 3 Jul 2017 16:53:35 +0800 Subject: [PATCH 099/314] fix set bugs for fInput and fGradInput in TimeDistributed layer code style From 4f92f58e2eb8ab37b56e4a19fe9e90748a35378a Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Tue, 4 Jul 2017 13:52:13 +0800 Subject: [PATCH 100/314] Checkin document pages (#1068) * checkin md doc files * add some docs * add more docs * change the folder location * fix unit test * fix typos * fix typo From ed95a2be0fc31c6a28fdaea6c876eae52f2905d4 Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Tue, 4 Jul 2017 21:21:26 +0800 Subject: [PATCH 101/314] revise hashCode to avoid null pointer From 5e852b278c4e826dd866c8147eac456552f51221 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Wed, 5 Jul 2017 16:21:50 +0800 Subject: [PATCH 102/314] Timedistributed add getParametersTable method (#1154) * override getParametersTable method * test getParametersTable From ffae7fa1baf756e5ed36a44274c799a537bbd6c2 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 6 Jul 2017 15:37:28 +0800 Subject: [PATCH 103/314] optimize GRU by pre-calculating three linear operations (#1150) * add preTopology for GRU optimize GRU * remove revisions in peephole * pass dropout test * add speed test * remove unnecessary variables; judge preTopology == Sequential * add comment From 2f25b706606e868e96c90193f3da8f4f05f16251 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 6 Jul 2017 16:55:12 +0800 Subject: [PATCH 104/314] Optimize LSTMPeephole by pre-calculating the input2hidden linear (#1158) * optimize LSTMPeephole * remove setname * test speed, and add class variable * rename timeDim to featDim in LSTMPeephole From 9f00c87f7dbaf4699cb09a245515e6fbecab789b Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Thu, 6 Jul 2017 18:47:07 +0800 Subject: [PATCH 105/314] Make SpatialFullConvolution declare with implict numeric type (#1128) * add test * remove unnecessary change in LinearSpec * Make SpatialFullConvolution declare with implict numeric type * rebase master * change SpatailaFullConvolution doc From 6d7f846894f6e8d724a6480c8ce1b9135e34a971 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 7 Jul 2017 10:12:11 +0800 Subject: [PATCH 106/314] Remove unnecessary varialbe in Recurrent object construtor (#1127) * add test * remove unnecessary change in LinearSpec * Remove unnecessary varialbe in Recurrent object construtor * pass the test From 7bb0d92a68658f8be0811ef7e85a1042b8b95a4b Mon Sep 17 00:00:00 2001 From: ding Date: Wed, 5 Jul 2017 11:56:46 -0400 Subject: [PATCH 107/314] support spatial convultion without bias From ea6ad9dd90856248c997cc786c359d062c1b8f24 Mon Sep 17 00:00:00 2001 From: ding Date: Thu, 6 Jul 2017 14:41:41 -0400 Subject: [PATCH 108/314] fix comments From 2b9d72caea4e32c470e553013a389751f456f920 Mon Sep 17 00:00:00 2001 From: ding Date: Mon, 26 Jun 2017 11:08:01 -0400 Subject: [PATCH 109/314] support convlstm and convolution without bias From b15db4690b7ddff003dc6ba046cdf013bd5a4b54 Mon Sep 17 00:00:00 2001 From: ding Date: Thu, 6 Jul 2017 18:16:37 -0400 Subject: [PATCH 110/314] fix comments From f4170c89efaa6f5343ecc5cc5fa49227ea14119e Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Mon, 10 Jul 2017 13:46:31 +0800 Subject: [PATCH 111/314] fix bugs in convLSTM From dfc59588546b13d7aedf597f4c314b7ec810cc03 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 11 Jul 2017 13:57:59 +0800 Subject: [PATCH 112/314] [Issue 864] save module weights, load (all / part) module weights (#1102) * load save module weights * add support for loading from another model * meet code review From 36685c32269794e5572c644ad683cd156daad34e Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Tue, 11 Jul 2017 14:35:31 +0800 Subject: [PATCH 113/314] Caffe persister (#1095) * add dynamic caffe loading support * refined with bug fix and more commnets * fix typo and bias check for convolution * renaming change * fix import reference * refine customized layer support * fix style check issue * resolve style issue * resolve conflict * caffe persister implementation * refinement with pr * rename loadcaffe name * fix typo * add docs From f1c300d7306027cf074fe3c9f7b495c47cdc204a Mon Sep 17 00:00:00 2001 From: JerryYanWan Date: Mon, 10 Jul 2017 10:59:42 +0800 Subject: [PATCH 114/314] RNN cells do not execute preTopology in Dropout mode From 01e8b8d2d8749033b13ceb6aada843f40df35076 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Tue, 11 Jul 2017 21:23:44 +0800 Subject: [PATCH 115/314] save/load model from s3 (#1136) From 381a6089fe054b16fa392faade2682f82807a0a9 Mon Sep 17 00:00:00 2001 From: ding Date: Mon, 10 Jul 2017 09:00:27 -0400 Subject: [PATCH 116/314] add optimation for convlstm From 3e90228f9d792474d66293078792094a7c71a18a Mon Sep 17 00:00:00 2001 From: ding Date: Mon, 10 Jul 2017 11:16:52 -0400 Subject: [PATCH 117/314] improve perf and add md From fe05012ea9864ead6e28a42e05cc23f6a2c14642 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 12 Jul 2017 14:28:31 +0800 Subject: [PATCH 118/314] Fix a bug in the MV layer (#1171) * add test * remove unnecessary change in LinearSpec * fix MV bugs From 9909c1e6306e6c8d97a7bda5446a0ee40f446d49 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 12 Jul 2017 15:51:00 +0800 Subject: [PATCH 119/314] Add 1D Convolution (TemporalConvolution) (#1180) * add test * remove unnecessary change in LinearSpec * finish Temporal Convolution * add python API * meet code review * meet code review * fix typos * fix typos * meet code review From 98717d6d8bc8844e253db7fe45d88100bc9fad70 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 12 Jul 2017 16:44:41 +0800 Subject: [PATCH 120/314] Fix maptable bug (#1214) * fix maptable bug * Add new line to end of maptable From 76411f929ab7639f626c1dd333aa4efac53ec57d Mon Sep 17 00:00:00 2001 From: jenniew Date: Fri, 7 Jul 2017 00:35:14 -0700 Subject: [PATCH 121/314] create localModule and update local predictor From 1373b7958754f422f9b4ef0fbfd4e46209a3f623 Mon Sep 17 00:00:00 2001 From: jenniew Date: Fri, 7 Jul 2017 00:44:54 -0700 Subject: [PATCH 122/314] update abstract module From 52fc2bcba5b05a58e7e16561b54efc4a5ab3fb80 Mon Sep 17 00:00:00 2001 From: jenniew Date: Mon, 10 Jul 2017 13:16:45 -0700 Subject: [PATCH 123/314] update local jvm with comments From 1033e735b71f5f89007f68f4906d18237fb3113e Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 12 Jul 2017 17:21:31 +0800 Subject: [PATCH 124/314] Make ClassNLL can process -1 label for padding, not update the model. (#1215) * add -1 for ClassNLL * fix typo * add tests From d733f90901bbf5ddc0627ec210f2976d5289b820 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Thu, 13 Jul 2017 06:23:49 +0800 Subject: [PATCH 125/314] Tree LSTM with Sentiment classification example (#1217) * finish treeLSTM with sentiment example * fix scala style * fix python error * fix a python bug * fix a typo * fix a typo From 5b665b64cf87d764cfdd872dd1c54c37407a2687 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Thu, 13 Jul 2017 12:18:09 +0800 Subject: [PATCH 126/314] fix inception and xavier (#1220) From df873d648c21fad416b98c059c114465eb686f5c Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 13 Jul 2017 15:17:57 +0800 Subject: [PATCH 127/314] BatchNormalization accepts default values (#1216) * BatchNormalization accepts default values * runningMean defined during run time * remove multiple constructor in BatchNormalization class From ee994fb14d8dc1c6ae5eb3367a1b3f2c3287078a Mon Sep 17 00:00:00 2001 From: ding Date: Thu, 13 Jul 2017 11:05:49 -0400 Subject: [PATCH 128/314] convert back convlstm optimization From dfdb03cc1720639e390f6d3aba2d3e8dd3861c3a Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 13 Jul 2017 23:29:36 -0700 Subject: [PATCH 129/314] fix concat when the dimension concatenated along is larger than 2 (#1230) * fix concat when dim >= 3 * fix backward * fix backward * changing back to tensor copy * meet code review From 1c2d630932ae6c7e7ef2213ce973e47c89de5ae4 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 18 Jul 2017 10:28:47 +0800 Subject: [PATCH 130/314] Fix the bug of JoinTable after calling clearState() (#1257) From 9c449c1ac7fe1466ced715879fe5eac26beffcd5 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Tue, 18 Jul 2017 12:13:43 +0800 Subject: [PATCH 131/314] optimize joinTable perf (#1248) From 7aa7bd16e8f9ae001b3f72ad06b1054c673f0aec Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 18 Jul 2017 15:10:31 +0800 Subject: [PATCH 132/314] add load caffe model doc, remove useless parameter, add python api (#1266) * add load caffe model doc, remove useless parameter, add python api * update doc * add save caffe * remove empty lines * fix python test * Update caffe-support.md * fix python * update doc From f658e378ce2bc67515bf519731d9b6450d2a3f55 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Tue, 18 Jul 2017 15:24:54 +0800 Subject: [PATCH 133/314] fix SpatialConvolution's hashcode and clearState will throw NullPointerException (#1268) From c91f5242ab4e92d0da745473873361084c481b25 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Wed, 19 Jul 2017 08:39:30 +0800 Subject: [PATCH 134/314] update shared convolution (#1144) update shared convolution with latest SpatialConvolution. Add wRegularize, bRegularizer, initWeight, initBias, initGradWeight, initGradBias, withBias From e16bfe45990209ee1fb76ee4579f17379369d109 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 19 Jul 2017 00:19:56 -0500 Subject: [PATCH 135/314] Graph doc (#1275) * Add Input docs * fix bugs * meet code review From 7772dece0378b78d53984d5b4007ca6004a2be4c Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Tue, 18 Jul 2017 22:45:28 -0700 Subject: [PATCH 136/314] Refine tensorflow doc (a follow up of pr #1253 to meet code review) (#1272) * meet code review * fix example * meet code review * refine title * make saveTF compatible with saveCaffe From b51c466961bca342c859368302f386ff2469410d Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Wed, 19 Jul 2017 20:28:29 +0800 Subject: [PATCH 137/314] add CosineDistanceCriterion (#1195) * add CosineDistanceCriterion * add cosineDistanceCriterion * add CosineDistanceCriterion python test * add python test * add blank line in pythonapi * revise APIDocs From dc0d0605055f2cfd006beeed6deae7deb4578641 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 20 Jul 2017 12:36:13 +0800 Subject: [PATCH 138/314] fix batchSize non-consistency bug (#1295) add unit test From ea3f8cf661d74e4b7139d9e58d366f44c32aa4c9 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 20 Jul 2017 00:39:06 -0500 Subject: [PATCH 139/314] Add scala tensorflow load/save examples (#1291) * add scala tf example * update * update * fix a small typo * fix unit test From 0ed660313aedc5819c154f3f6e76316c0cb0a93a Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 20 Jul 2017 14:43:13 +0800 Subject: [PATCH 140/314] Recurrent: fix clearState (#1306) From 13ea525adc21ca5a6f1194d2194e399251c9870c Mon Sep 17 00:00:00 2001 From: dding3 Date: Mon, 24 Jul 2017 20:59:26 -0700 Subject: [PATCH 141/314] remove unnecessary contiguous check in volumetricconvolution (#1357) remove unnecessary contiguous check in volumetricconvolution From 543a3762af05d7f7c9b1d7027ddd40bba4c6fb25 Mon Sep 17 00:00:00 2001 From: dding3 Date: Tue, 25 Jul 2017 22:17:34 -0700 Subject: [PATCH 142/314] Convlstm3d (#1361) * support convlstm3d From 9757dbd1a73ba1c18e26840918d6297db67fa6ac Mon Sep 17 00:00:00 2001 From: Xianyan Date: Wed, 26 Jul 2017 15:35:22 +0800 Subject: [PATCH 143/314] Support global average pooling (#1365) * add global pooling * update doc and python * update doc From 61d443e51cc28a9893dc5c5ebb7f05ba97d3995f Mon Sep 17 00:00:00 2001 From: Xianyan Date: Wed, 26 Jul 2017 16:18:23 +0800 Subject: [PATCH 144/314] add SpatialWithinChannelLRN, update the mapping of caffe lrn (#1364) * add SpatialWithinChannelLRN * add python * fix map error * fix python * meet code review * update doc From 4bf88dacb5d4ace09e1dfbded230dc1fc478099a Mon Sep 17 00:00:00 2001 From: dding3 Date: Wed, 26 Jul 2017 22:37:11 -0700 Subject: [PATCH 145/314] support get final state and cell (#1369) * support get final state and cell From 4e5e25662aae256458384d7dd5768409315bcf0e Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Thu, 27 Jul 2017 14:49:26 +0800 Subject: [PATCH 146/314] Fix temporalconv (#1359) * add dynamic caffe loading support * refined with bug fix and more commnets * fix typo and bias check for convolution * renaming change * fix import reference * refine customized layer support * fix style check issue * resolve style issue * resolve conflict * add missing method From 1713f68127fd39ca26446fda50918fc1c13ca7fe Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 27 Jul 2017 16:45:20 +0800 Subject: [PATCH 147/314] add regulaizer to volumetricconvolution (#1371) * add regulaizer to volumetricconvolution * fix name * update doc * update doc From fc4fd2005b5db912a7819f2d8b7d312641a4fe0e Mon Sep 17 00:00:00 2001 From: dding3 Date: Thu, 27 Jul 2017 23:10:48 -0700 Subject: [PATCH 148/314] add regularizer for convlstmpeephole3d (#1395) * add regularizer for convlstmpeephole3d * rename convlstmpeephole2d back to convlstmpeephole From 2aa51233a8fe9d757e65b2526205ed8c4cfd3f46 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Wed, 2 Aug 2017 14:27:18 +0800 Subject: [PATCH 149/314] Check target size in ClassNll (#1420) From 3e610768be1f14afbf87ab91fdc2f9ecb4f71bec Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Thu, 3 Aug 2017 11:04:21 +0800 Subject: [PATCH 150/314] Add 3D de-convolution (#1401) * finish volumetric full convolution * add documentation * reset core * meet code review * add serial version uid * meet code review From fcdb4f0355d8ea4e8dc1bc3dc022767e9c88adb7 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 7 Aug 2017 01:58:46 -0700 Subject: [PATCH 151/314] add nhwc im2col and col2im kernel (#1419) * add nhwc im2col and col2im kernel * address comments From 38251491a6911f35e22e792871b8a66584624c5a Mon Sep 17 00:00:00 2001 From: dding3 Date: Thu, 10 Aug 2017 20:34:53 -0700 Subject: [PATCH 152/314] Set hiddenstate (#1439) * support set/get state in recurrent From 86371f390797b3110ad89077f89e8249b22fbead Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Fri, 11 Aug 2017 13:56:37 +0800 Subject: [PATCH 153/314] LSTM to Graph (#1428) From ce22aae616c08d96473804e5d341cc3925a72c62 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Fri, 11 Aug 2017 13:57:10 +0800 Subject: [PATCH 154/314] SimpleRNN to Graph (#1431) * SimpleRNN to Graph * move i2h, h2h to buildGraph in SimpleRNN From 914216b2581af5a36ab6d526dff99b50f9b5ba13 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Fri, 11 Aug 2017 13:57:28 +0800 Subject: [PATCH 155/314] GRU to Graph (#1429) * GRU to Graph * set withBias = false in gru h2g From 59aa377c8bef3354ff20a6cfcf79d0b91f5b9321 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Fri, 11 Aug 2017 15:07:57 +0800 Subject: [PATCH 156/314] revise LSTMPeephole to Graph (#1430) From 7ac4bbfcc5c207200f6acd3811a1581079e7bddb Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Mon, 14 Aug 2017 11:05:38 +0800 Subject: [PATCH 157/314] add funcional model definition (#1200) * add inception and lenet graph * add resnet,autoencoder,alexnet graph * fix graph unit test * add vgg graph and test * add inception_v2 graph * add inception-v1 graph * add graph model example * change JoinTable nInputDims From 78e5653d2aac7f78a805fc3fea4f1472e05492dc Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Mon, 14 Aug 2017 14:05:41 +0800 Subject: [PATCH 158/314] fix copyStatus in TimeDistributed layer (#1451) From 7d8612d8065f499c99b16f86f628ed79ba006431 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Mon, 14 Aug 2017 15:25:24 +0800 Subject: [PATCH 159/314] optimize logsoftmax (#1424) * optimize logsoftmax * use ev instead of MKL * add clearState * revise precision of InceptionSpec and EvaluatorSpec using MKL * revise GoogleNet_v1 test precision * change buffer name From 76007c1a307b3d99738187a4f9c7021e8ccb6f3d Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 14 Aug 2017 05:55:26 -0700 Subject: [PATCH 160/314] support nhwc format in SpatialConvolution (#1434) * support nhwc format in SpatialConvolution * add python api and doc * meet code review * fix bug and add tests * meet code review * fix bug From 6baef96116c8afe989ebe8b867f4f3506af37f3e Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Wed, 16 Aug 2017 10:51:55 +0800 Subject: [PATCH 161/314] remove set operation (#1464) From 058a1cac5eee3096be84090579bc2b59a0d759c8 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Wed, 16 Aug 2017 15:54:39 +0800 Subject: [PATCH 162/314] add mkl.tanh (#1450) * add mkl.tanh * change core * chang tanh From c358955cef423a23bb899cb127c2d33514ee822f Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 16 Aug 2017 17:12:24 +0800 Subject: [PATCH 163/314] Add abstract Operation, Add, Substract, Multiply and Divide operations (#1462) * finsh Operation, Add, Substract, Multiply and Divide operations * meet code review * make acc not be supported * meet code review * make documentation better * rename package object * rename package object From 0f372cd19c8e36a28dcb495c7ada41e7ba6836b8 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Wed, 16 Aug 2017 19:06:11 -0700 Subject: [PATCH 164/314] support same padding in maxpooling and avgpooling (#1458) * support same padding * add doc * refactor * add doc From 377d90cd4dfc0d6b9f0f41fb314eb9445ee5d2e4 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Thu, 17 Aug 2017 13:43:41 +0800 Subject: [PATCH 165/314] optimize Sigmoid (#1471) From 2a29808ffcf78f81e32427a4fa0ddeec9bee3f7a Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Thu, 17 Aug 2017 16:24:28 +0800 Subject: [PATCH 166/314] add getTimes in recurrent and TimeDistributed layer (#1470) * add getTimes * add getTimes for all Cells * addTime method in cell * add appendTimes method * remove nullException when frequently calling setTimes and getTimes in recurrent layer * fix bugs in calculating backwardSum * add getTimes for Graph * revise ConcatTable * remove add in calculating backwardTime From 69fe0e2722c293add643e1a21df0e5f7527a0ecd Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Fri, 18 Aug 2017 17:06:11 +0800 Subject: [PATCH 167/314] support layerwise serialization (#1472) * support layerwise serialization * support layerwise serialization * add unit tests * remove commented code * fix style issue * change invoke api * add new api * resolve conflict * fix typo * refinement per review * refinement per review * refinement From 2f75079c81bb90ddcba15479bb4d1e1085d1ed37 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Sun, 20 Aug 2017 19:51:48 -0700 Subject: [PATCH 168/314] Support NHWC format in SpatialMaxPooling (#1444) * support nhwc in SpatialMaxPooling * add tests * fix bug and meet code review * add python api and doc From 56d4d427125d81bc5edc87dfe923efc46208d1ef Mon Sep 17 00:00:00 2001 From: Xianyan Date: Wed, 23 Aug 2017 10:44:11 +0800 Subject: [PATCH 169/314] [Issue #1033 #450] Disable backpropagation for general layers (#1181) * add propagateback and generate backward graph new unit test pass unit test except resnet pass unit test pass support rebuild support multiple inputs fix updateGradInput and acc fix unit test rename to trainable, add API to set trainable at graph unit test except inception and resnet pass unit test pass code clean fix conflicts add unit test for setTrainable, setFreeze, unFreeze use input buffer add getSubmodule in AbstractModule resolve conflicts update isInputNode update comment update setTrainable fix fix mixture of graph and container add setFreeze python API update python and add doc meet code review * meet code review * fix unit test * meet code review * fix code style * fix inception unit test * meet code review * meet code review * rm cloneNode * resolve conflicts * fix reset * meet code review * meet code review * fix From f1b8f790cd0e05d974c9a46fcd4ead2367a6cde0 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Wed, 23 Aug 2017 13:47:54 +0800 Subject: [PATCH 170/314] fix inplace layer clearstate (#1496) * fix inplace layer clearstate * meet code review * fix bug of LeakyReLU when inplace is true From 4586cfa84bf68294acf32a7973e79a2f326e64ed Mon Sep 17 00:00:00 2001 From: Hawkwood <2041829103@qq.com> Date: Thu, 24 Aug 2017 11:22:24 +0800 Subject: [PATCH 171/314] Give Detailed information of nn exception (#1482) * Give Detailed information of nn exception * modify some changes * bug fix1 From b87cf0f95df823f3202dc064ea0ab3cf38c59ae3 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 24 Aug 2017 00:27:39 -0700 Subject: [PATCH 172/314] support NHWC format in SpatialAveragePooling (#1488) * support NHWC format in SpatialAveragePooling * fix style * meet code review From e71a4dcf5538a1417330eb8ebc5136907ab7a860 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Fri, 25 Aug 2017 08:25:40 +0800 Subject: [PATCH 173/314] Serialization - version control & python API (#1491) * add version control * version support * add python api * add version control proto * add unit test * change function name * rename function api * per comments * refinement to make hierachy more cleaner * backward compatibility From 9415d4f3d0f520e20c59c23548ab70d01378b96a Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Fri, 25 Aug 2017 11:09:27 +0800 Subject: [PATCH 174/314] Recurrent supports BatchNormalization (#1176) * addPreprocessInputLayer to support BatchNormalization * revise bnorm to hiddensize * add comments * add BatchNormalization between preTopology and topology * add BatchNormalization in Recurrent * recurrent support batchnormalization * revise sequential to identity in throw new Error * fix serialize * add serializeBigDLModule * return Unit in doSerializeModule From 370291a66bc8247e6060b559fc8aea5863463b93 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Mon, 28 Aug 2017 09:19:07 +0800 Subject: [PATCH 175/314] change tanh backward (#1502) From daf7344ef5e7d67df2c01d33b043cd14bcdacb11 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Mon, 28 Aug 2017 12:59:18 +0800 Subject: [PATCH 176/314] add BatchParams for BiRecurrent layer (#1512) From a9ba423241a6a5876c3f302cc3cbf6724a59da05 Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Mon, 28 Aug 2017 14:58:45 +0800 Subject: [PATCH 177/314] BifurcateSplitTable (#1513) * add BifurcateSplitTable * createBifurcateSplitTable * add buffer * clearState From 170f72c5341e35bb96390b7a195eadcb1c362aac Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Tue, 29 Aug 2017 08:40:36 +0800 Subject: [PATCH 178/314] add isSplitInput in BiRecurrent (#1514) * add isCloneInput in BiRecurrent * rename isCloneInput to isSplitInput * rename test class From 391d3feaf0526fe6271c7a35c65640b6add410b7 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Tue, 29 Aug 2017 01:18:04 -0700 Subject: [PATCH 179/314] Remove the code dealing with translating tensorflow NHWC format to BigDL NCHW format (#1506) * rm process dims code * fix rnn and nchw test case From d345ad10a4010c5313948e7e2cf46dae1fe23f6b Mon Sep 17 00:00:00 2001 From: Yan Wan Date: Wed, 30 Aug 2017 11:20:08 +0800 Subject: [PATCH 180/314] add isInputWithBias and isHiddenWithBias for RNNCell (#1515) * add isInputWithBias and isHiddenWithBias for RNNCell * default affine = true in BatchNormParams in recurrent layer * add serializableSpec for recurrent From 19737f3a6e9120db25e6cf248328d90a2832e85d Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Fri, 1 Sep 2017 15:22:10 +0800 Subject: [PATCH 181/314] serialization - add custom data conversion support (#1511) * add custom data conversion support * fix issue * add missing change * add missing change From 33ab8e8fe5ff07d777118029926985a67f1fe553 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 1 Sep 2017 15:27:13 +0800 Subject: [PATCH 182/314] Add LogicalNot, LogicalOr and LogicalAnd (#1495) * add boolean, short, int, long, string, char to tensor support * implement and test LogicalOr and LogicalAnd operations * add logical not * change a typo * refine logicalOr and logicalAnd * fix scala code style error From 23dfc7a222800ca3c74bc3b392a8775a440e38a7 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 1 Sep 2017 15:29:46 +0800 Subject: [PATCH 183/314] 1D Max Pooling (#1436) * finish 1D pooling * finish documentation * meet code review * finish 1d pooling * meet code review * fix scala style errors From 6fedfc5ca1e98a3a04ab1d6e44c1e57db748832e Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 1 Sep 2017 16:34:21 +0800 Subject: [PATCH 184/314] Add Squeeze, Prod, Sum, Reshape, Identity, ReLU (#1477) * finsh Operation, Add, Substract, Multiply and Divide operations * meet code review * make acc not be supported * meet code review * add some operations * add prodcut * add default prod method in UndefinedTensorNumeric class From 34baa6cfc4b468428383add8f01612e7f35d4a40 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Mon, 4 Sep 2017 11:19:57 +0800 Subject: [PATCH 185/314] Allow module output multiple tensor in Graph (#1507) * Allow module output multiple tensor in Graph * fix failed unit test * meet code review * meet code review * more unit test * add more unit tests * fix unit tests * fix style check * fix style check and unit test * meet code review * support nested output * meet code review From 6f368cfa35d1e8dcf8af9775afeab779893a4fa7 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 4 Sep 2017 14:38:04 +0800 Subject: [PATCH 186/314] support loading a subset of tf graph and support control dependency in graph (#1499) * support loading a subset of tf graph * support load model with specified inputs * add control dependency support in Graph * fix file reader conflict * fix bug * add some comments * fix control dep bug * fix doc From 810059432e8dfa7280e64efaeb3d2bdd4179abde Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 6 Sep 2017 14:45:12 +0800 Subject: [PATCH 187/314] Make layer suppor different input and output numeric type and Add Equal Operation as an example (#1520) * make a layer support different type * finish equal and make layer suppor different type * make code clear * move zipwith to tensor * fix conflict * fix scala style error * change an implementation for make scala2.10 compile successfully From 416dbb4e4b5e956648f37f18b565446ec70a641d Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Thu, 7 Sep 2017 09:52:58 +0800 Subject: [PATCH 188/314] Add Greater and Less operations (#1525) * finish greater and less * fix scala style From d213b4df9870d7514e5681c37d1996149325c904 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Tue, 12 Sep 2017 11:32:51 +0800 Subject: [PATCH 189/314] Display bigdl model in tensorboard (#1545) * Display bigdl model in tensorboard * add document, python API and meet code review. * meet code review * meet code review From e16f24698058b5f31cde58e91e4eedcc1359df82 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Tue, 12 Sep 2017 13:39:27 +0800 Subject: [PATCH 190/314] Fix operation logical related bugs (#1549) * Support control flow * Decouple input/output tensor numeric type from the module numeric type(parameter type). * FIx unit tests * while loop api refactor * fix unit test * make breeze version configurable as spark 2.1 bump breeze version * Add scheduler to exclude layers * remove control-flow change * remove Schduler and ControlOps * fix broken unit test From b4919497fcd03df3a4dc109e20f07c4b88d17fb9 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 14 Sep 2017 13:21:05 +0800 Subject: [PATCH 191/314] Support switch and merge operation (#1555) * Support control flow * remove loop * fix unit test * fix unit test From 88e2e4583b467dc09fcba92e8ad350247d3d9762 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 15 Sep 2017 10:34:58 +0800 Subject: [PATCH 192/314] Add Floor L2Loss RandomUniform Rank MatMul SoftMax operations (#1533) * make a layer support different type * finish equal and make layer suppor different type * fix conflict * add Floor L2Loss RandomUniform Rank MatMul SoftMax * add TruncatedNormal * fit new code * make jenkens pass From e7e78e5ea937b9d64dbd2da530ec2e541c3e3501 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 15 Sep 2017 15:41:49 +0800 Subject: [PATCH 193/314] Implement and test Conv2d operation (#1539) * finish conv2d * add conv2d test * fit new code * meet code review From 1b5ad05658c40558fab36d2c12043b86e264b5b8 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Fri, 15 Sep 2017 17:23:04 +0800 Subject: [PATCH 194/314] Fix failed tensorflow unit test (#1563) * fix tf unit test * fix unit test * meet code review From 29626d863a58b8c27a53a2029d03a1659c4c231a Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 18 Sep 2017 13:02:57 +0800 Subject: [PATCH 195/314] support multi output in graph follow up (#1564) * support multi-output follow up * meet code review From 4d0aaf96eab401a06ebc6db9078a2ba2856c0b4a Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Tue, 19 Sep 2017 13:34:00 +0800 Subject: [PATCH 196/314] optimize copy in recurrent (#1556) * optimize copy in recurrent * add more unit test * select optimize * meet review From 42fafc3ba0679d1799d359081b043114be2be998 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Tue, 19 Sep 2017 18:51:59 +0800 Subject: [PATCH 197/314] support different type in tensorflow loader (#1569) * support different type in tensorflow * meet code review * meet code review * fix tests From 4cb832c141143585aade656fc7eb82dfc09be2c2 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Wed, 20 Sep 2017 10:02:05 +0800 Subject: [PATCH 198/314] Add assert, image decoders, parseExample operation (#1559) * add some ops * add parse example * style * add variable * add png decoder * add raw decoder * remove resources * fix ops * test sig * remove useless code * remove import * fix style * fix conflict * fix api From cc21eaf0cf087fd191f1d973f673abea104093ea Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Wed, 20 Sep 2017 21:47:37 +0800 Subject: [PATCH 199/314] Add slice and onehot (#1547) * Add slice and onehot * refactor the code * fit new code * fix a typo * meet code review * fix test failed From 956593ad0244eb44103cd6f6bb27b8e5b701605f Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Thu, 21 Sep 2017 10:16:16 +0800 Subject: [PATCH 200/314] Handle null array (#1571) * fix type comparision * handle null value array * fix type diff From f7ea6315e50008ed31a1eac42087ada0d31ebb13 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 21 Sep 2017 11:53:21 +0800 Subject: [PATCH 201/314] Support load tf unet (#1575) * add deconv2d, resizebilinear and support broadcast in add * fix test * fix style error and meet code review * fix python failure * fix unit test From aa31da47974f02889b99df734d1a92d9d448e1ec Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Thu, 21 Sep 2017 16:02:03 +0800 Subject: [PATCH 202/314] fix Logsoftmax infinity error (#1576) From 08161f96ed9842910f4a35fe339d6c34605fac35 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Mon, 25 Sep 2017 15:22:15 +0800 Subject: [PATCH 203/314] Add Assign, Cast, ExpandDims, MaxPool operations (#1562) * add assign * Add Cast and ExpandDims * meet code review * meet code reiew * meet code review * meet code review * meet code review * add comment to NumericWildcard type * try to fix an AbstractMethod error in Cast method * fix test failure From e66d7f6e8f293121b8b95ce658cd59dc560ae300 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Mon, 25 Sep 2017 15:33:33 +0800 Subject: [PATCH 204/314] Add Realdiv, BiasAdd and Pad (#1579) * Add Pad * refactor biasAdd From 423027c89dd9b7d788c7c4256a0d0959ae3754a0 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 25 Sep 2017 16:23:09 +0800 Subject: [PATCH 205/314] Support Tensorflow Preprocessing (#1572) * fix tensor bug * preprocessing * test * fix tests * refine * fix tests * fix style * fix tests * meet code review * meet code review * meet code review * add doc From 1e1466b33d6194a42e3e9d28cd48cc39de93b22b Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Tue, 26 Sep 2017 14:48:56 +0800 Subject: [PATCH 206/314] refactor predictor (#1582) * refactor predictor * predictClass share model output memory * refactor repeatMemory to shareBuffer From ce4f777a39c3bafbc8c92b33817eeace786f8579 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 27 Sep 2017 16:54:11 +0800 Subject: [PATCH 207/314] Refine tensorflow loader code strcuture (#1593) * revise tensorflow loaders * refine more layers * fix sytle issue * fix style issue * more ops * meet code review * fix style issue From 46a7195bbe17f06be95e5ae66f2f40dc0b0cd423 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 27 Sep 2017 17:13:42 +0800 Subject: [PATCH 208/314] 1. throw exception when ConcatTable contains no submodule (#1598) 2. rescursive remove stopGradient nodes From da17fa758634430b598ef738e8423d2e33f75125 Mon Sep 17 00:00:00 2001 From: dding3 Date: Wed, 27 Sep 2017 13:42:18 -0400 Subject: [PATCH 209/314] Support feed prior prediction back into recurrent (#1476) * Implement RecurrentDecoder * add ut for lstm with recurrent decoder From 34a3b0f5fa7d62d0c57e3b02025fbac4be632aef Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Fri, 29 Sep 2017 13:23:31 +0800 Subject: [PATCH 210/314] Refine scheduler (#1607) don't execute relevant nodes in forward/backward fix a bug when handle const node From b63bea22a60766ff2d91c64c029191f087e31241 Mon Sep 17 00:00:00 2001 From: jenniew Date: Sun, 13 Aug 2017 14:06:33 -0700 Subject: [PATCH 211/314] fix MapTable issue From 99da09f0ed2fc8efb87b2af45e932b2bc229f5fb Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Mon, 9 Oct 2017 15:14:31 +0800 Subject: [PATCH 212/314] Tensor types support (#1630) * refinemnt to support more tensor types * refinemnt * refinemnt * support object sharing * support shared memory * remove unused type * refinement From 2a232786911c0550b589365bf90127239468b2e2 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Mon, 9 Oct 2017 04:33:18 -0400 Subject: [PATCH 213/314] feat: quantization methods (#1610) Quantization object support: 1. quantize a value with max and min. 2. quantize an array 3. quantize a Tensor[Float] And for test, there're relative dequantize methods. From 7a5dfa350b629e944d99b8d4b35a45325546c719 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Mon, 9 Oct 2017 08:20:01 -0400 Subject: [PATCH 214/314] feat: quantized tensor support (#1611) * feat: quantized tensor support QuantizedTensor inherits from Tensor with supporting part of methods. The memory is allocated from native code, so it only matains a pointer `getNativeStorage`. Note, We don't support automatically `free` method. Anyone using this method should think about the memory lacking, especially for memory optimization. test: quantized tensor tests * fix: test case failed on spark 1.6 From af0aaccf686f51829e1df3a1b1a260b0e88e1fc4 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Tue, 10 Oct 2017 10:26:06 +0800 Subject: [PATCH 215/314] support graph (#1636) From 8bf29dc119210b3fa01954063eb955de78135400 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Tue, 10 Oct 2017 04:21:36 -0400 Subject: [PATCH 216/314] feat: serilization for quantized modules (#1613) * feat: serilization for quantized modules All quantized modules extending QuantModule, which has an empty Tensor for gradient. And the object mixes QuantSerializer for protobuf supporting. * refactor: serialization api changes From 3470d5e20540781ab2462e0c85bf552ed615c41c Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Tue, 10 Oct 2017 06:02:05 -0400 Subject: [PATCH 217/314] feat: quantized convolution (#1614) * feat: quantized convolution * fix: serialization api changes From 1f5c0a5701fe54e3f9118524d83d921526f5bae4 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Tue, 10 Oct 2017 06:45:17 -0400 Subject: [PATCH 218/314] feat: quantized linear (#1615) * feat: quantized linear * fix: serialization api changes From 05283d3b60ca3be68c87910be51a9f37bc8e44d5 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Tue, 10 Oct 2017 07:43:28 -0400 Subject: [PATCH 219/314] feat: quantized dilated conv (#1617) * feat: quantized dilated conv * fix: serialization api changes From 1ddcaa63fc964125528774b792e6cf0680e0b70a Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Wed, 11 Oct 2017 03:16:48 -0400 Subject: [PATCH 220/314] feat: quantize a whole graph/modules (#1618) * feat: quantize a whole graph/modules * feat: python supports * fix: delete unusage From 2c4b9d83db9fb81d119e8db09d9d5164a0238b70 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Thu, 12 Oct 2017 00:36:14 +0800 Subject: [PATCH 221/314] Add tile operation (#1646) * Add tile operation * reset core * remove unnecessary comments From 39113e9a2afbaefdc011b36a8cbcf4389e7a06bd Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Thu, 12 Oct 2017 13:16:43 +0800 Subject: [PATCH 222/314] add Variational Auto Encoder (#1432) * vae * add python api and unit tests * add docs From 5c66b437be1b0f8bcdce1734c372c2ce8664a43b Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 12 Oct 2017 15:49:35 +0800 Subject: [PATCH 223/314] serveral fix for preprocessing (#1609) * serveral fix for preprocessing * load all * fix tests * more fixes * meet code review * fix style * fix style * fix tests * fix tests * rebase and fix api * refine test case * fix tests * add transpose * fix tests * fix expandims * fix serializer * fix doc From e3b1518c41bff993c686454435fc1b352e4ee2b0 Mon Sep 17 00:00:00 2001 From: dding3 Date: Thu, 12 Oct 2017 22:03:22 -0400 Subject: [PATCH 224/314] RecurrentDecoder enhancement (#1619) * RecurrentDecoder refactor From 7cbfdb88d462bef28de047650aa07ba2f024bb96 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Fri, 13 Oct 2017 12:51:24 +0800 Subject: [PATCH 225/314] throw clear exception in lookuptable (#1586) From cd9c7b6cf7aedc945d6da440a3ba7a9368d25427 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Fri, 13 Oct 2017 16:44:26 +0800 Subject: [PATCH 226/314] SparseLinear SparseJoinTable DenseToSparse (#1652) * SparseLinear SparseJoinTable DenseToSparse * Python api * add DenseToSparseSpec * update to upstream * add some method * meet code review * fix python unit test * fix python unit test From f4ccc9f75470525332825570bcd7939e32a5eb18 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Mon, 16 Oct 2017 11:28:35 +0800 Subject: [PATCH 227/314] move model freeze to AbstractModule (#1647) * move model freeze to container * update * update freeze API * refactor freeze API * meet code review and add unit test * fix doc * meet code review From ec031c72a0592c5625ddfce606d2215d7acc8eee Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Mon, 16 Oct 2017 17:29:30 +0800 Subject: [PATCH 228/314] Add BroadcastGradientArgs (#1658) * finish BroadcastGradientArgs * remove breeze * pass the tests * refine the function * fix scala style * while instead of for From d48acf9ab9aaf2435eb25618ba6d5c679c8c748d Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Wed, 18 Oct 2017 09:56:11 +0800 Subject: [PATCH 229/314] Fix duplicated msg proto (#1659) * fix protobuf duplicate msgs * fix test cases From e72e2fae8203c41865b44860381ef18c0b7427f6 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 19 Oct 2017 09:56:14 +0800 Subject: [PATCH 230/314] add more layers and operations (#1678) * add more layers and operations * fix compile error * fix compile error From a1508bc7f6b1683864dd3d27b6b71c1dc5a54bf6 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 19 Oct 2017 10:08:24 +0800 Subject: [PATCH 231/314] fix decode raw and StridedSlice and Transpose serilization issue (#1673) From c58cf4b34836ac1f0be74c0bf8faba687755e1a3 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Thu, 19 Oct 2017 14:24:36 +0800 Subject: [PATCH 232/314] add ptb model example (#1672) * add ptb model example * meet review From 0cebe9d5f0be4c0edbe0d3b786ddbe908c2cb41b Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 19 Oct 2017 14:35:17 +0800 Subject: [PATCH 233/314] Support load more tensorflow operations (#1656) * Support load more tf operation * refine session * fix style issue * meet code reivew From e8244efb6c0dc6779f2d7afa3ab33a9f24a93cc3 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Fri, 20 Oct 2017 04:33:06 -0400 Subject: [PATCH 234/314] fix: make the length of weight and gradWeight in quantized conv same length (#1692) * fix: the different length of weight and gradWeight in quantized conv is very confused. So make them same length. * fix: return grad From a5ae91c6d2acc898ce8157a95f83073b7f0937e8 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Mon, 23 Oct 2017 10:45:34 +0800 Subject: [PATCH 235/314] add toGraph (#1189) * add toGraph * refactor toGraph * add getEndNodes for ParallelTable, MapTable * add unit test From f2018c66144453ea02afc552fd1a32a4daa6ee40 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 25 Oct 2017 20:36:48 +0800 Subject: [PATCH 236/314] fix shape output tensor type from T to Int (#1737) From 2d593c46a2d6fa42f1eab895d8840fd5731837d0 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Fri, 27 Oct 2017 11:24:40 +0800 Subject: [PATCH 237/314] Clone graph keep their original prevs nodes and next nodes sequence (#1763) * Clone graph keep their original prevs nodes and next nodes sequence * add comments * update * simplify table merge From 19c068e3c23c357638d898cdbec02e1fa1ca9183 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Fri, 27 Oct 2017 14:19:38 +0800 Subject: [PATCH 238/314] MapTable: Add module back in clearstate (#1778) * MapTable: Add module back in clearstate * add unit test From 6656112ec872ff7e3c3e503b248d73e0d65be4db Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Tue, 31 Oct 2017 10:00:11 +0800 Subject: [PATCH 239/314] MSRA filler (#1802) From a9a48ef47f6580befa64d23eb1311d3b764d0db8 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Tue, 31 Oct 2017 10:01:25 +0800 Subject: [PATCH 240/314] support argmax operation (#1714) From fd60a000f6f44113d30b3e2511bb8cc30f2295fa Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Wed, 1 Nov 2017 15:37:25 +0800 Subject: [PATCH 241/314] Caffe layer wise unit test (#1810) * add run caffe jobs * fix style * fix typo * chmod * add test info * refinement * fix import * refinement * refinement to get exit status * fix typo * fix path * add layers unit tests * add python interface * fix style From debbc93d92ad2a8ce01c1c17bbf80c26102c4e40 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Thu, 2 Nov 2017 10:28:48 +0800 Subject: [PATCH 242/314] optimize MSECriterion and Adam (#1701) From 0bbcd78854861e70d44cb5797a7f3944bc656bde Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 2 Nov 2017 15:01:19 +0800 Subject: [PATCH 243/314] Support more tensorflow operations (#1818) * add more operations * add more ops * fix style issue * fix style issue * remove some jdk8 pai From 7fb2dadba7073da8eb04c044e8b6425949b4695b Mon Sep 17 00:00:00 2001 From: Hawkwood <2041829103@qq.com> Date: Fri, 3 Nov 2017 08:43:55 +0800 Subject: [PATCH 244/314] Fix cmul (#1822) * fix CMul * 1697 issue * add exception capture and modify unit test From 9b24d126c38752514522103c38c800356438b1e2 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Fri, 3 Nov 2017 09:55:51 +0800 Subject: [PATCH 245/314] Add some operations (#1816) * add some ops * add SquareDifference * squaredifference * add SquaredDifference test * add tests * add abs * fix style * fix style * fix style * fix tests * fix style * add div * meet code review * meet code review * meet code review From de2a9e8dac8c6b66363d1d6e2eac0130e18259b2 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Fri, 3 Nov 2017 16:47:38 +0800 Subject: [PATCH 246/314] Fix SmoothL1Criterion and SoftmaxWithCriterion (#1830) From b6451207711b8fa38d8f2a5b9b28b9eef82d4e13 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 6 Nov 2017 14:27:20 +0800 Subject: [PATCH 247/314] Add square related ops (#1823) * add square related ops * use another type parameter * fix error message From b777c42cc9ee85d4ef9fd89ca69f280232693bff Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 6 Nov 2017 19:37:28 +0800 Subject: [PATCH 248/314] Add Maximum, Minimum, BatchMatMul (#1828) * add ops * fix style From d1e162fb3cdd8fd978f105c69764957fff645901 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Thu, 9 Nov 2017 08:40:58 +0800 Subject: [PATCH 249/314] Tf ParseExample add serialization support (#1852) * update bigdl download url * fix (#1719) * fix jtensor (#1722) * [Backport 0.3 PR1728]Bump script (#1729) * checkin bump-version script * add comments * add serialization support From 45a62d2a12b30aefebdfc012af3c3098776a22f6 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Fri, 10 Nov 2017 13:28:07 +0800 Subject: [PATCH 250/314] time counting for paralleltable's forward/backward (#1861) From a9112a8799c46007f7074e3d0698c67f3ff7cca2 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Fri, 10 Nov 2017 15:48:11 +0800 Subject: [PATCH 251/314] Support load more tensorflow operations (#1858) * add more ops * checkin reduce * add All * support all and any * add operation Round * support floordiv and truncatediv * support FloorMode/Mod/TruncateMod * add exp/expm1 * add InTopK * add rint * remove a JDK8 api * fix style issue * fix a bug in floordiv From 46b6b15c5544341f16cf9d10c250bc59cca75560 Mon Sep 17 00:00:00 2001 From: dding3 Date: Fri, 10 Nov 2017 15:16:15 -0500 Subject: [PATCH 252/314] support SAME padding in 3d conv and allows user config padding size in convlstm and convlstm3d (#1862) * support SAME padding in 3d conv * allows user config padding size in convlstm and convlstm3d From 9ffa8d08d116fe1a876cadfb4523213df43934d8 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Tue, 14 Nov 2017 09:48:48 +0800 Subject: [PATCH 253/314] fix loading ds2 module issue (#1867) * fix loading ds2 module issue * refinemnt * refinement per review From 7c17e0ac3dae1e6dfa5bd3dff3fd1775594fcf80 Mon Sep 17 00:00:00 2001 From: Wang Yanzhang Date: Tue, 7 Nov 2017 03:56:00 -0500 Subject: [PATCH 254/314] refator: update initialize method for serialization From bd8097b14560c778381ff1e70c82585ec1cb9a80 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Tue, 14 Nov 2017 12:45:08 +0800 Subject: [PATCH 255/314] Support keras squared_hinge and sparse_categorial_crossentropy (#1865) * add squared hinge loss * classNLL support prob as input * add python api * meet code review * fix python tests * more doc * more doc From a9b3cacd6aff6f89697cde98f99868ad01fd5082 Mon Sep 17 00:00:00 2001 From: "li,zhichao" Date: Wed, 15 Nov 2017 09:15:24 -0600 Subject: [PATCH 256/314] Support keras model loading (#1827) * keras load model and api wrapper * fix batchnorm * fix averagepooling * remove function wrapper and clean code * add unit test * remove backend * fix and clean * update * separate unittest * x * remove -n * ignore application * separate test * pass parameter * try to fix test * disable shape checking * ignore From 1ff4bbc17a9fcd6a5a59a1212c4a28880702770c Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Fri, 17 Nov 2017 00:09:00 -0600 Subject: [PATCH 257/314] Keras support - add Hard-sigmoid layer (#1897) * add hard sigmoid * add python * add python api * fix typo * add doc From f574783f7709237ac096ac86a2b5642d35670b71 Mon Sep 17 00:00:00 2001 From: Kai Huang Date: Tue, 21 Nov 2017 11:09:43 +0800 Subject: [PATCH 258/314] Add more keras layers and test (#1891) * add, update and fix * add hard_sigmoid * fix python3.5 lambda in permute * fix syntax for python3.5 * fix pooling and conv layers * fix syntax * fix permute for py35 From b5ef25947cc83e71dde2c526656e69fc7b5596b2 Mon Sep 17 00:00:00 2001 From: Kai Huang Date: Mon, 27 Nov 2017 09:06:20 +0800 Subject: [PATCH 259/314] Add serialization test for CAveTable and VolumetricAveragePooling (#1926) * add serialization uid * add ut for serialization From 8654df46c61dc809a1712e897844294c7cc2a2d8 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Mon, 27 Nov 2017 02:04:59 -0600 Subject: [PATCH 260/314] fix graph missing generatebackward (#1877) From c5cb592d63929df172466e8b1ce1d698b471f536 Mon Sep 17 00:00:00 2001 From: Kai Huang Date: Mon, 27 Nov 2017 16:54:13 +0800 Subject: [PATCH 261/314] add activation for recurrent layers (#1909) From a808c0aa6f9b382f41352c2c7c2886c385c02d04 Mon Sep 17 00:00:00 2001 From: Shane Huang Date: Mon, 27 Nov 2017 17:47:46 +0800 Subject: [PATCH 262/314] add GaussianDropout, GaussianNoise layers (#1894) * add GaussianDropout, GaussianNoise layers, unit tests, python wrapper, module serializer unit test and docs From ab60cab53f2460b242a1a1dbdb7d09ff00dcbce7 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Mon, 27 Nov 2017 19:09:19 -0600 Subject: [PATCH 263/314] Keras support - add cosine proximity criterion (#1902) * keras support to add CosineProximityCriterion * fix criterion * add unit test * fix typo * add comment From b233affcdf09dbe35e581c199203bd5c96e33e31 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 28 Nov 2017 19:05:11 +0800 Subject: [PATCH 264/314] Add keras highway (#1913) * Add highway * Fix unit test * Add python support * fix ut * add module seri * fix ut * Add API doc * Add Regularizer From 5f8cbc7900d128f6da17039ce51b7d3bf6fdc5be Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 30 Nov 2017 13:26:32 +0800 Subject: [PATCH 265/314] Add Upsample3d (#1911) * add * Add upsample3d * add SerialVersionUID * add python * Fix ut * add saveModule test * add updample3d serializer and markdown doc * meet code review * meet code review * fix ut * fix ut * resolve conflicts From 4671809646dddaf30438ce9afc84978c874bda08 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Mon, 4 Dec 2017 16:31:47 +0800 Subject: [PATCH 266/314] Add an API to get node from graph model with given name (#1871) * get node from graph * add unit test * add Graph.node api in python From 02d186814f580e2ea0545e908b04e27fe1ad4c21 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Tue, 5 Dec 2017 15:07:59 +0800 Subject: [PATCH 267/314] Support NHWC for LRN and batch normalization, and optimize batch normalization speed (#1936) * add LRN and LRNGrad * add LRNGrad * support bn with NHWC * support bn with NHWC * add batch norm grad * support tf bnv2/bngradv2 * fix uts * fix uts * fix ut * fix ut * fix ut * fix ut * fix style issue * fix ut * fix ut * fix ut * update python api and docs * meet code review * meet code review * fix ut * fix ut * fix ut From 3c6e08f4f776f64ed1436336a9567b9f254553cc Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Tue, 5 Dec 2017 02:05:32 -0600 Subject: [PATCH 268/314] Pb support bigmodel (#1976) * support new format * add unit test * add unit test * refinement * fix typo * fix unit test From e7ca47e837d7088495e2332da72d2029bdde0ba1 Mon Sep 17 00:00:00 2001 From: Yanzhang Wang Date: Tue, 5 Dec 2017 17:42:49 +0800 Subject: [PATCH 269/314] feat: SReLU and ActivityRegularization support (#1917) * feat: SReLU and ActivityRegularization support * fix: update doc * fix: refactor the weights and gradWeights of SReLU * fix: serialization for srelu * fix: backward computing errors * fix: ActivityRegularization to layer * fix: change test location * fix: add python api * fix: shared axes have not consider enough From c0202a79f9552e9e23c1a02f020707ce96a545ba Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Tue, 5 Dec 2017 19:57:44 -0600 Subject: [PATCH 270/314] add support for multiple tensornumerics (#1984) From 6a4d228113f86dcd5a50d54e6f05e6b232e63b83 Mon Sep 17 00:00:00 2001 From: dding3 Date: Wed, 6 Dec 2017 14:17:03 -0500 Subject: [PATCH 271/314] Support Keras masking and maxoutdense (#1918) * support keras Masking layer * support keras maxoutdense From 62731ce67768de56e6edac167ba13d078b8571f1 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 7 Dec 2017 13:22:57 +0800 Subject: [PATCH 272/314] fix output allocation bug (#1982) * fix output allocation bug * fix test * fix rebase * fix test * fix ut From 33bc824adf867345e3dabc5654f7d084d81dec1a Mon Sep 17 00:00:00 2001 From: dding3 Date: Thu, 7 Dec 2017 15:15:46 -0500 Subject: [PATCH 273/314] Add multiple rnn cell and support get/setHiddenState in Recurrent(#1591) * add multicell * add get/setHiddenStates From 4619ab5390d12463f21ff0d2669182c0be8a32ae Mon Sep 17 00:00:00 2001 From: Kai Huang Date: Fri, 8 Dec 2017 18:13:02 +0800 Subject: [PATCH 274/314] Map newly added keras layers on python side (#1990) * update * add maxoutdense * add masking * add srelu * highway activation as string; update docs * code clean From f48c0ed46734c7cfd6012647e08f9460d8354a2e Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Mon, 11 Dec 2017 15:18:41 +0800 Subject: [PATCH 275/314] add Ser support to all tf related layers (#2002) * add Ser support to all tf related layers * fix style From 569ee7aa97852633f82789c6e113e892c291bdf6 Mon Sep 17 00:00:00 2001 From: zhangxiaoli73 <380761639@qq.com> Date: Mon, 11 Dec 2017 16:49:59 +0800 Subject: [PATCH 276/314] add MeanAbsolutePercentageCriterion & MeanSquaredLogarithmicCriterion (#1916) * add MeanAbsolutePercentageCriterion & MeanSquaredLogarithmicCriterion * add docs * meet pr comments * meet pr comments * exchange keras input target * update readme From 01be6ce1a40cd0daf3268578543f3f917914c75d Mon Sep 17 00:00:00 2001 From: Yuhao Yang Date: Mon, 11 Dec 2017 18:21:56 -0800 Subject: [PATCH 277/314] [keras][loss] kullback_leibler_divergence and poisson (#1938) * add kdl and poisson * add extra unittest * poisson update * kullback * update kld * kld update * style fix * add transient * add extra ut * init with null * resolve conflict From 579f55b5a31857509449479439ec3562d6624e94 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Tue, 12 Dec 2017 15:06:53 +0800 Subject: [PATCH 278/314] Support SperableConv2D and one hot crossentropy (#1944) * Support SperableConv2D and one hot crossentropy * fix compile error * add python wrapper * add new ut * add ut * fix unit test * fix ut * refine depthwiseConv2D * fix seperable conv2d unit tests * support DepthWiseConv2dBackpropInput and DepthWiseConv2DBackpropFilter * refine the code * add more test * fix failed test and style issue * fix inconsistant API with keras1 * add serialization unit test * fix failed unit test From 231ff8b5e8fd858e06ff9285f25a8863ce20b5d9 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 12 Dec 2017 16:54:13 +0800 Subject: [PATCH 279/314] Add object detection related layers for model zoo and keras support (#2015) * Add object detection model zoo support * Add unit test * Add more doc and python wrapper * revert some and add createProposal * Fix ut From bcb808382478a148dea0490511e273715469c499 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Tue, 12 Dec 2017 17:21:25 +0800 Subject: [PATCH 280/314] support serialization unittest auto-check (#2011) * support serialization auto check * fix typo * merge master From bc4758b2705b45aa88e8e13f886a439e608c12d4 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 12 Dec 2017 18:04:07 +0800 Subject: [PATCH 281/314] Add image frame predict (#1972) * Add VisionPredictor * share convolution * inception example can work * move predictImage To Predictor add matToTensor, remove postPorcessor add python Add more doc * meet code review and fix unit test * fix test * remove unnecessary ClassTag in ImageFrame From 46365ab6b340f36082bd0b30d34e1719789aa9f3 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Wed, 13 Dec 2017 10:29:18 +0800 Subject: [PATCH 282/314] Add serialize unit test for new add layer (#2021) * Add serialize unit test for new add layer * add PriorBox serializer From 5a392462143bf3107477f3d86fe89acc402c53fa Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Wed, 13 Dec 2017 16:13:58 +0800 Subject: [PATCH 283/314] UpSampling1D UpSampling2D (#1928) From ef12a19dd480546f4d35c5b7eb5c1830bb489073 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Wed, 13 Dec 2017 17:51:12 +0800 Subject: [PATCH 284/314] Fix some issues found when save bigdl model to tensorflow format file (#1997) * fix BN save tf issue when scale and offset is null; fix conv2d save tf issue when bias is null; support save LRN layer to tf format * fix get padding info not correct when ceilMode=true and pad=0 * support save Scale in tf format * refind some code * fix unit test From 8961d576d1557871fbc2d0401a075609af2bd150 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Wed, 13 Dec 2017 18:01:06 +0800 Subject: [PATCH 285/314] add more ops (#1847) From ff346bc2eb1accaf9b9478a4e84aa8e2fd1079ab Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Wed, 13 Dec 2017 18:55:48 +0800 Subject: [PATCH 286/314] Added Dilation2D, Dilation2DBackpropFilter, Dilation2DBackpropInput (#1853) * add ops * fix style * add more docs * address comments * fix test * add serilization tests From 5d6dbb4331cbfaf69c213eebe99e206b93cd5fed Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 14 Dec 2017 09:46:22 +0800 Subject: [PATCH 287/314] Add conv3d related operations (#1915) * add conv3d * backpropinput * fix conv3d * add backpropinputv2 * add backpropfilter and refactor * fix style * address comments * add serialization tests * fix tests From 79e95dbf8c30e96a88a727deeaa4a56f4feb7917 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Thu, 14 Dec 2017 11:01:35 +0800 Subject: [PATCH 288/314] add tensorflow operation SegmentSum (#1956) From 41b9cc70b1c9140294be534c09d1713c3bc748a6 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Thu, 14 Dec 2017 14:09:49 +0800 Subject: [PATCH 289/314] Fix 1923: Use static graph execution when there's no control operation in graph container (#2008) * checkin dynamic graph and static graph * fix compile error * refine graph * meet code review * fix serialize/load test * fix failed unit test * add unit test for dynamic graph * merge duplicate code in graph * fix generateback issue * fix python test * fix style issue * meet code review * use array to cache output instead of hashmap * meet code review * fix a confused part mentioned in code review From c10ab1564e0f6a4e92da975c2116a25fb9232582 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 14 Dec 2017 17:56:32 +0800 Subject: [PATCH 290/314] Add local model predictImage (#2026) * Add local model predictImage From e9e6f2e4ff1d770a1bdd38b905e2fef394cb1d36 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Fri, 15 Dec 2017 09:37:50 +0800 Subject: [PATCH 291/314] get large model from executor (#2017) * fix of getModel * meet code review * meet code review * fix ut * delete copy status From b597017810b384d152b58bc90505c4880a95098d Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Fri, 15 Dec 2017 12:53:17 +0800 Subject: [PATCH 292/314] Concat check input size check (#2034) * add dim check * fix typo * per review * per review * per review From 31d1519bbf095f55f66f473031ca5907000a8c2b Mon Sep 17 00:00:00 2001 From: Yuhao Yang Date: Thu, 14 Dec 2017 23:29:51 -0800 Subject: [PATCH 293/314] Add Cropping2D and Cropping3D (#1992) * add cropping * serialization test * equal and hash * style fix * python doc From 9da2247f5afd96c01ccd5abc2217d06624c6a82a Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 15 Dec 2017 17:23:55 +0800 Subject: [PATCH 294/314] Refine textclassifier with data without headers and refactor to temporal convolution (#1707) * refine text classifier * add more comments to TemporalMaxPooling * update document * refine textclassification scala code From e5a9fb1c526e1f2d4def30f399312994c3725288 Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Fri, 15 Dec 2017 19:26:09 +0800 Subject: [PATCH 295/314] Add keras SpatialDropout1D, SpatialDropout2D, SpatialDropout3D (#1898) * add comments * meet review * add python code * add documentation * add serialzation test From 921feefcc0ccf4aa8fd4c1f001c085c880e89b67 Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Mon, 18 Dec 2017 09:35:05 +0800 Subject: [PATCH 296/314] Lookup table for multivalue (#2016) Layer LookupTableSparse From 29212b1f3a8ce036e908b3e6d7a2937bf3fc5bff Mon Sep 17 00:00:00 2001 From: Yao Zhang Date: Mon, 18 Dec 2017 12:16:35 +0800 Subject: [PATCH 297/314] Add locallyconnected2d layer (#2037) * finish locallyconnected2d * add locallyconnected2d * add serializer test * fix python test failed * meet code review * change spaticalConvolutionSpec * remove a test From ec716d6d8567f8d15051535bf3c5d40143ce1573 Mon Sep 17 00:00:00 2001 From: Ian Wong Date: Mon, 18 Dec 2017 14:12:18 +0800 Subject: [PATCH 298/314] fix null pointer exception in bn when affine is false (#2045) * fix null pointer exception in bn when affine is false * meet code review From f64b92766b26a61d6c44fc19d3ced6f39ba57497 Mon Sep 17 00:00:00 2001 From: Guoqiong Song Date: Mon, 18 Dec 2017 12:54:21 -0800 Subject: [PATCH 299/314] Locallyconnected1D (#1964) * add LocallyConnected1D, address review comments, resolve conflicts, add serializer test From 978a117cf18d8262d01b665fd1803b6cd25d3671 Mon Sep 17 00:00:00 2001 From: Kai Huang Date: Thu, 21 Dec 2017 11:17:15 +0800 Subject: [PATCH 300/314] Map Keras LocallyConnected1D (#2060) * add locallyconnected1d * fix style From 46d64ceec7f27d40a6337afe14676b352c555dd6 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 21 Dec 2017 15:30:50 +0800 Subject: [PATCH 301/314] Fix KLDCriterion forward (#2078) * fix KLDCriterion * add python doc * add unit tests * add api doc From 9f7d6c0a060d67d804cbc137046e3793e250848c Mon Sep 17 00:00:00 2001 From: Xianyan Date: Thu, 28 Dec 2017 10:08:18 +0800 Subject: [PATCH 302/314] optimize localpredictor for small dataset (#2110) * optimize localpredictor for small dataset * Add featurePaddingParam * Add comments * refactor LocalPredictor.predict * add log tmp * fix ut From 331cc262c91d891a6f3b54ef2f1a4671721fd732 Mon Sep 17 00:00:00 2001 From: Jerry Wu Date: Thu, 28 Dec 2017 13:51:35 +0800 Subject: [PATCH 303/314] graph ser to support control node (#2121) * graph ser to support control node * add unit test * refinement * refinement From f33af50bb9e0d92485f5edd711e765bc8d985536 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 28 Dec 2017 15:07:13 +0800 Subject: [PATCH 304/314] Refine KLDCriterion (#2113) * optimize KLDCriterion * address comments * fix tests From 417bba4deacaff8f48aeab9a0f6f6d39dc23359d Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Thu, 28 Dec 2017 17:53:31 +0800 Subject: [PATCH 305/314] fix resize bilinear (#2124) * fix resize bilinear * fix style * fix tests From 66435fb69e3bae62dc21f9912e77e8c85b84c5bf Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Tue, 2 Jan 2018 13:52:38 +0800 Subject: [PATCH 306/314] Add TransformerCriterion (#2097) * add TransformerCriterion * fix style * fix python tests * meet review From 21dad26cbff59d9352333cb0e92224394f27eede Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Fri, 5 Jan 2018 15:55:23 +0800 Subject: [PATCH 307/314] Wrap layers that requires two ClassTags into TensorModuleWrapper and move to tf package (#2122) * add TensorModuleToOperation * fix tests * address comments * fix test * rebase master * fix style From 2ec0cd65aa4490d40b1f5c6f200c83075b3f4b66 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Wed, 10 Jan 2018 13:33:02 +0800 Subject: [PATCH 308/314] fix spatialfullconvolution bug (#2175) From 4b40cece51ea562de1cec3f6d44295326348321f Mon Sep 17 00:00:00 2001 From: Yuhao Yang Date: Sun, 14 Jan 2018 21:55:12 -0800 Subject: [PATCH 309/314] udpate equal (#2033) From 15c9662cd5ff900d36b6900af18b7b94aa3ba529 Mon Sep 17 00:00:00 2001 From: Xu Xiao Date: Tue, 16 Jan 2018 10:45:13 +0800 Subject: [PATCH 310/314] Module DenseToSparse support backward disable (or not) setting (#2189) * Module DenseToSparse support backward disable (or not) setting * change isBackward to propagateBack * change isBackward to propagateBack From 39ec1ef0e3bbf151ed7ce0ef8affdd22b14e59b3 Mon Sep 17 00:00:00 2001 From: Xianyan Date: Tue, 16 Jan 2018 18:35:34 +0800 Subject: [PATCH 311/314] Add default zeroGradParameters to AbstractModule (#2177) * Fix optimizer NaN in SSD * Add zeroGradient * Add default zeroGradParameters * revert From 2f4a7732fb0bf499199ba2f6ca83038996714a2f Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Fri, 19 Jan 2018 15:21:35 +0800 Subject: [PATCH 312/314] fix recurrent (#2202) From 0c7f994ba8dc7b858b3fbc7a59ab26c822c7db7b Mon Sep 17 00:00:00 2001 From: Xin Qiu Date: Mon, 22 Jan 2018 09:25:35 +0800 Subject: [PATCH 313/314] fix leakyrelu (#2206) From 03e4f06d6fde6fd4cb3b898682be79bfe927ad62 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Mon, 22 Jan 2018 13:11:17 +0800 Subject: [PATCH 314/314] Add NCHW support for resizebilinear (#2205) * add nchw support for resizebilinear * fix style * fixed indent * test cover more cases * remove double code and python interface * add doc