TastyPancakes/Flux.jl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2,589 Commits 104 Branches 42 Tags 5.8 MiB
86d6555269
Commit Graph

1668 Commits

Author SHA1 Message Date
Mike J Innes
b05e755068 rm jit from cuda 2018-04-15 15:08:58 +01:00
tejank10
5cc681317a added stride for pooling in tracker 2018-04-15 15:07:04 +01:00
tejank10
f6097d58d6 Scalar pad/stride for Conv constructor 2018-04-15 12:15:41 +05:30
Mike Innes
9d7164f15f we'll do this differently 2018-04-14 02:09:35 +01:00
tejank10
65847bb745 moved epsilon into sqrt 2018-04-04 15:25:20 +05:30
tejank10
3ead662987 Update rule fixed 2018-04-04 15:18:44 +05:30
CarloLucibello
b415333233 fix reshape 2018-04-02 16:09:57 -04:00
tejank10
ea9b5471fa NADAM optimizer 2018-04-03 01:27:22 +05:30
Brad Safnuk
b9a66c679d Fix error in initialization of σ. 2018-03-22 22:20:21 -04:00
Brad Safnuk
35299d4621 Fix type instability when loading onto a gpu. 
Also fixes Issue #216.
 2018-03-22 21:32:32 -04:00
Mike J Innes
4320738d87 fix 2018-03-21 11:25:47 +00:00
Mike Innes
1c5f8e3534 ndims for shapes 2018-03-16 14:42:08 +00:00
Brad Safnuk
db2d9efb72 Update BatchNorm documentation 2018-03-15 21:59:38 -04:00
Brad Safnuk
6653ec86d9 Allow multidimensional inputs to batchnorm. 
Can be used in conjunction with convolutional layers, in addition
to dense layers, with the same api.
 2018-03-15 21:48:59 -04:00
Mike J Innes
e931552f7d
Merge pull request #200 from chengchingwen/repmat 
implement `back` of `repmat`
 2018-03-15 15:18:48 +00:00
Mike J Innes
5d7edb5aaa
Merge pull request #197 from chengchingwen/master 
Implement `prod` for `TrackedArray`
 2018-03-15 15:17:24 +00:00
boathit
2ec37790be eliminate ambiguity 2018-03-13 10:50:56 +08:00
boathit
ff2caf032c eliminate ambiguous 2018-03-12 22:48:16 +08:00
Mike J Innes
9ccbac8b80 jit gpu support 2018-03-07 19:18:27 +00:00
chengchingwen
43af3895b0 change prod implementation to avoid small xs 2018-03-07 21:03:13 +08:00
chengchingwen
c00f7f850f implement back of repmat 2018-03-07 20:43:59 +08:00
chengchingwen
7c721475c6 add gradient check for prod and fix dims in back(::typeof(prod),...) 2018-03-07 16:24:44 +08:00
Mike Innes
d21c313ea7 tweaks 2018-03-06 19:58:47 +00:00
Mike Innes
36baa7ec2c convnet primitives 2018-03-06 19:58:05 +00:00
Mike Innes
0802b4d5cf closes #198 2018-03-06 16:56:01 +00:00
Elliot Saba
6445295318 Better download detection 2018-03-06 08:45:45 -08:00
Elliot Saba
19f691d342 Use cache.julialang.org to store ML models 
It's annoying that when third party servers go down our tests break.
Let's at least make sure that if our tests break due to server outages
it's our fault.
 2018-03-06 08:03:21 -08:00
Mike Innes
3babeeb440 scalar hashing 2018-03-06 13:49:05 +00:00
chengchingwen
86d782a5ce implement prod for TrackedArray 2018-03-06 18:01:19 +08:00
Mike Innes
c95a97f6ae make epochs available 2018-03-06 03:01:56 +00:00
Mike Innes
432b9c3222 loadparams! 2018-03-06 02:45:31 +00:00
Mike Innes
65ed95190a fix 2018-03-05 23:44:25 +00:00
Mike Innes
bfd6a4c0ec cleaner interrupts 2018-03-05 23:05:45 +00:00
Mike Innes
5153cde847 move epochs 2018-03-05 22:56:22 +00:00
Mike J Innes
662439c164 closes #177 2018-03-05 17:24:46 +00:00
Elliot Saba
36295799ee Add permutedims() for tracked arrays 2018-03-02 10:22:28 -08:00
Mike J Innes
8019f789f8 use normal log 2018-03-01 16:35:49 +00:00
Mike J Innes
ac57fc3c26 use @ fix in a few places 2018-03-01 16:31:20 +00:00
Mike J Innes
c2fea2acf6 revert this 2018-02-28 23:06:53 +00:00
Mike J Innes
2eb38eedbf update gpu api 2018-02-28 22:51:08 +00:00
Mike J Innes
ccef9f4dd4 jit softmax 2018-02-28 22:07:35 +00:00
Mike J Innes
7606b1a399 single-batch convolution 2018-02-28 14:25:32 +00:00
Mike J Innes
6bdc2b37a9 inline call 2018-02-28 13:47:14 +00:00
Mike J Innes
a401f08cda compile layers 2018-02-27 22:40:51 +00:00
Mike J Innes
5a32976cbf basic compile step 2018-02-27 21:43:41 +00:00
Mike J Innes
bdb8aae107 move cache logic 2018-02-27 21:41:03 +00:00
Mike J Innes
2c74976602 more general 2018-02-27 01:25:40 +00:00
Mike J Innes
466b5c501a cpu/gpu conveniences 2018-02-26 23:10:59 +00:00
Mike J Innes
15d1d3256b conv api updates 2018-02-26 22:43:07 +00:00
Mike J Innes
54919b8dca rm deprecation 2018-02-22 00:23:02 +00:00
Mike J Innes
491785a681 ignore state in mapleaves 2018-02-22 00:22:51 +00:00
Mike J Innes
ec65e2cec7 fix printing 2018-02-22 00:21:48 +00:00
Mike J Innes
af2e6b7e1d fix 2018-02-22 00:15:38 +00:00
Mike J Innes
99b739cf00 fixes #176 2018-02-21 23:21:20 +00:00
Mike J Innes
e3b4b16e01
Merge pull request #178 from schmrlng/pull-request/e6f55641 
Convert OneHot CuArrays to dense CuArrays before passing to CUDNN methods
 2018-02-21 22:34:11 +00:00
Mike J Innes
6bdd283fbd no longer necessary 2018-02-21 22:29:31 +00:00
Iblis Lin
043fedde3c
introduce Reexport 
- Reexporting NNlib

fix #180
 2018-02-21 16:55:20 +08:00
Ed Schmerling
e6f556411a Convert OneHot CuArrays to dense CuArrays before passing to CUDNN methods 2018-02-19 17:32:15 -08:00
Mike J Innes
4035745f6e may help numerical tests 2018-02-19 12:51:02 +00:00
Mike J Innes
989adcdc7d gpu fix 2018-02-17 12:41:53 +00:00
Mike J Innes
11511982a4 numerical stability 2018-02-17 11:56:03 +00:00
Mike J Innes
e5791bc5f6 frequencies utility 2018-02-17 11:19:51 +00:00
Mike J Innes
e3b31b9b87
Merge pull request #169 from jessebett/jessechanges 
Reshape with Tuple Dimensions and Kronecker Product
 2018-02-16 14:16:42 +00:00
Mike J Innes
60f21d3ff2 don't override base method 2018-02-16 14:15:40 +00:00
Mike J Innes
5e861101f3 epochs util 2018-02-16 11:17:57 +00:00
Mike J Innes
7aa6854c64 more correct 2018-02-16 00:06:15 +00:00
Mike J Innes
ee3784964e fix for external modules 2018-02-15 22:27:00 +00:00
Mike J Innes
63862c2324 easier initialisation with weights 2018-02-15 20:52:29 +00:00
Mike J Innes
01c31e7fcc conv bias 2018-02-15 20:15:41 +00:00
Mike J Innes
bdd07a8bc6 fix 2018-02-14 22:34:11 +00:00
Mike J Innes
1b8b1cd7b1 check params by identity 2018-02-14 21:00:50 +00:00
Mike J Innes
5ea0ef6764 tracker fix 2018-02-13 16:15:36 +00:00
Mike J Innes
1baa7227e3 reorganise batches 2018-02-13 16:05:07 +00:00
Mike J Innes
34217b1fa2 Merge branch 'treebank' 2018-02-13 15:44:27 +00:00
Mike J Innes
49584fb72b rm logsigmoid 2018-02-13 14:52:29 +00:00
Mike J Innes
2f29733888 Merge branch 'master' into HEAD 2018-02-13 14:45:37 +00:00
Mike J Innes
8432d8db06 batchnorm fix 2018-02-13 14:02:35 +00:00
Mike J Innes
820cd3ae42 fixes #164 2018-02-13 13:31:35 +00:00
Mike J Innes
066cb45a38 remove old accuracy fn 2018-02-13 11:12:21 +00:00
Mike J Innes
236edbffec fixes #111 2018-02-13 10:20:38 +00:00
Mike J Innes
f22cfb5b43 re-enable printf 2018-02-12 15:05:09 +00:00
Mike J Innes
334ae9e1cb fixes #171 2018-02-12 12:31:15 +00:00
Mike J Innes
0b3c02fe8d document regularisation, fixes #160 2018-02-09 19:00:26 +00:00
Mike J Innes
0e0057b0c4 basics 2018-02-09 13:51:07 +00:00
jessebett
f84ee8eab0 reshape with tupled dimensions and kronecker product 2018-02-08 14:27:57 -05:00
Mike J Innes
70fbbf48fa humble beginnings of compiler 2018-02-08 18:11:26 +00:00
Mike J Innes
fc157a8c59 TrackedNumber -> TrackedReal 2018-02-08 17:18:40 +00:00
Mike J Innes
d1c56ca768 number fix 2018-02-08 17:04:48 +00:00
Mike J Innes
0f7a1ec022 test params funct 2018-02-08 16:13:20 +00:00
Mike J Innes
961de2ba44
Merge pull request #161 from FluxML/curnn 
WIP: CUDNN RNNs
 2018-02-08 13:06:52 +00:00
Iblis Lin
f7fdfbe3a9 fix params 2018-02-08 12:56:10 +00:00
Mike J Innes
fcbdc49d6b fix reserve usage 2018-02-08 10:27:26 +00:00
Mike J Innes
bc452fcd81 rewrite tests 2018-02-08 02:37:55 +00:00
Mike J Innes
d592f4e327 batch support 2018-02-08 01:45:48 +00:00
Mike J Innes
b8f148b012 hook up backward passes 2018-02-08 00:49:39 +00:00
Mike J Innes
a1d1930097 Merge branch 'master' into curnn 2018-02-07 23:23:02 +00:00
Mike J Innes
4511936a87 fixes #116 2018-02-07 23:21:04 +00:00
Mike J Innes
0ac924e8e1 fixups 2018-02-07 22:52:46 +00:00
Mike J Innes
39f7f8fdf3 tracked tuples 2018-02-07 22:21:42 +00:00
Mike J Innes
79e4e25fea seperate number type 2018-02-07 20:39:36 +00:00
Mike J Innes
282889970d seperate tracking infrastructure from array wrapper 2018-02-07 17:43:25 +00:00
Mike J Innes
30b3437c56 backward passes 2018-02-06 18:56:17 +00:00
Mike J Innes
f866fbe575 nullable c refactor 2018-02-06 15:01:48 +00:00
Mike J Innes
07e1b1e0a9 avoid val 2018-02-06 12:44:18 +00:00
boathit
7e37a96c6f Register back! for logsigmoid and implement (logit)binarycrossentropy 2018-02-06 19:36:16 +08:00
boathit
6e65789828 Register back! for logsigmoid and implement (logit)binarycrossentropy 2018-02-06 19:32:46 +08:00
Mike J Innes
a4bf5936b0 diagm 2018-02-05 18:29:35 +00:00
Mike J Innes
2fec75005d
Merge pull request #123 from GenaBitu/cat-fix 
Added vcat for multiple TrackedVectors
 2018-02-05 18:10:48 +00:00
Mike J Innes
47cebab26e test multiple inputs/dims 2018-02-05 18:09:54 +00:00
Mike J Innes
2a2475a9c2 get tracker graph 2018-02-05 17:40:07 +00:00
Mike J Innes
14086b8c2d train forward pass 2018-02-02 17:48:08 +00:00
Mike J Innes
9a6fcf057b hook up interface 2018-02-02 16:42:18 +00:00
Mike J Innes
b1c5786012 Merge branch 'master' into curnn 2018-02-02 15:56:44 +00:00
Mike J Innes
49e1e78f67 make data/value available 2018-02-02 15:56:04 +00:00
Mike J Innes
0f1e7b5578 update rnn structure 2018-02-01 20:57:39 +00:00
Mike J Innes
106502a75d typo 2018-01-31 21:57:04 +00:00
Mike J Innes
af3ccf85ff coagulate gates 2018-01-31 16:56:27 +00:00
Mike J Innes
4bfb603da6 gru forward 2018-01-31 13:46:55 +00:00
Mike J Innes
b1bb05403c basic forward pass 2018-01-30 18:18:37 +00:00
Mike J Innes
0b886507dc param offsets 2018-01-30 14:43:39 +00:00
Mike J Innes
af0c5523ff rnnTrainingReserveSize 2018-01-30 14:43:39 +00:00
Mike J Innes
3fb83d642d rnnWorkspaceSize 2018-01-30 14:43:39 +00:00
Mike J Innes
6b4e114d5d rnnParamSize 2018-01-30 14:43:39 +00:00
Mike J Innes
ee6c3e18a9 basic RNNDesc 2018-01-30 14:43:39 +00:00
Mike J Innes
842bf03051 typo 2018-01-30 14:43:05 +00:00
Mike J Innes
0c9549c469 rm lazy 2018-01-24 13:28:52 +00:00
Mike J Innes
5118ef9163 remove batching work for now 2018-01-24 13:12:38 +00:00
boathit
374d7a5f1e Registering backward function for logsoftmax 2018-01-21 15:20:59 +08:00
Mike J Innes
72eabde373 load data 2018-01-17 16:39:55 +00:00
Mike J Innes
bd57359535 docstrings 2018-01-17 16:12:12 +00:00
Mike J Innes
8cca7accf2 mnist 2018-01-17 15:55:37 +00:00
Mike J Innes
4207fb98f2 basic GPU tests 2018-01-16 17:58:14 +00:00
GenaBitu
bc8a32bc56
Merge branch 'master' into cat-fix 2018-01-16 11:01:31 +01:00
Mike J Innes
1beb30e19a closes #118 2018-01-15 17:00:47 +00:00
Mike J Innes
8f8589a7f4 fix initialisation 2018-01-10 14:11:52 +00:00
Mike J Innes
b44237468e Merge branch 'master' into gru 2018-01-10 13:59:33 +00:00
Mike J Innes
805cb9178f fixes #146 2018-01-10 12:48:50 +00:00
Mehul Tikekar
2fef799109 fix typo in conv.jl (fixes #133) 2018-01-08 16:46:58 -05:00
Mike J Innes
468f641f66 use Adapt 2018-01-08 16:34:22 +00:00
Mike J Innes
98b362729d pool padding 2017-12-18 18:18:14 +00:00
Mike J Innes
e3577d759c conv docs 2017-12-18 18:05:48 +00:00
Mike J Innes
269d8f36b9 conv padding 2017-12-18 18:05:38 +00:00
Mike J Innes
51f93d9f0e conv polish 2017-12-15 16:24:45 +00:00
Mike J Innes
386eafc443 reshape 2017-12-15 16:18:16 +00:00
Mike J Innes
73ae25289d remove old util 2017-12-15 16:18:01 +00:00
Mike J Innes
6890a61587 todo 2017-12-15 16:17:45 +00:00
Mike J Innes
9b833a4345 more onehot indexing 2017-12-15 16:17:39 +00:00
Mike J Innes
9d0dd9fb7e layer wip 2017-12-15 13:22:57 +00:00
Mike J Innes
0bf22dfb8e pool gradients 2017-12-15 02:29:14 +00:00
Mike J Innes
d949b31aa5 conv gradient 2017-12-15 02:24:32 +00:00
Mike J Innes
5b97d2ba04 closes #127 2017-12-13 18:24:56 +00:00
Mike J Innes
95d1287455 Merge branch 'master' into jacobian 2017-12-13 17:06:23 +00:00
Mike J Innes
27d896943e
Merge pull request #120 from staticfloat/sf/dense_initialization 
Better default initialization for Dense layers
 2017-12-13 16:18:02 +00:00
Mike J Innes
e3a688e706 use kwarg 2017-12-13 15:27:15 +00:00
Mike J Innes
128725cefd Merge branch 'master' into sf/weighted_crossentropy 2017-12-13 15:14:47 +00:00
Mike J Innes
29787eba45 fixes #114 2017-12-12 17:23:15 +00:00
Mike J Innes
b7b6c975bc fixes #110 2017-12-12 17:07:39 +00:00
Mike J Innes
403cc26327 Merge branch 'master' into gru 2017-12-12 16:54:00 +00:00
Mike J Innes
86097e76fd tweak batchnorm example 2017-12-08 19:34:34 +00:00
Mike J Innes
6f997e798a Merge branch 'master' into batchnorm 2017-12-08 19:31:50 +00:00
Mike J Innes
1d916c81b5 Merge branch 'master' into HEAD 2017-12-08 18:31:55 +00:00
Mike J Innes
24a6569589 Merge branch 'master' into amsgrad 2017-12-08 18:20:53 +00:00
Mike J Innes
f82dbf4798 Merge branch 'master' into HEAD 2017-12-08 17:00:31 +00:00
Mike J Innes
951c21366a fix regex 2017-12-08 16:42:30 +00:00
GenaBitu
7e51418679
Added back for multi-parameter vcat 2017-12-08 16:10:09 +01:00
baggepinnen
385dee9d16 Add jacobian function 2017-12-08 14:46:12 +01:00
GenaBitu
41f3eedc39
Proper multi-variable vcat 2017-12-07 17:50:18 +01:00
Elliot Saba
41446d547f Add weighted_crossentropy for imbalanced classification problems 2017-12-05 17:09:05 -08:00
Elliot Saba
c59b820bed Add glorot (Xavier) initialization 
Set default `Dense` and `RNN` inits to `glorot_uniform()` for `W`, `zeros` for `b`.
 2017-12-05 14:24:48 -08:00
GenaBitu
62b3600eca
Merge branch 'master' into cat-fix 2017-12-05 11:13:29 +01:00
baggepinnen
41febee9c1 Export and indent 2017-12-04 09:34:27 +01:00
baggepinnen
36001d085a Implement AMSGrad optimiser 2017-12-04 09:17:05 +01:00
Mike J Innes
cab235a578 gpu compat 2017-11-30 13:51:31 +00:00
Mike J Innes
19039f4881 export sigmoid 2017-11-30 13:37:38 +00:00
Mike J Innes
2d33f19346 onehot unk arg 2017-11-29 16:45:50 +00:00
baggepinnen
fa718c7475 Implement Gated Recurrent Unit 2017-11-24 14:33:06 +01:00
CarloLucibello
13b934c250 improve optimizers 2017-11-24 12:12:20 +01:00
Mike J Innes
dc1f08a709
Merge pull request #98 from FluxML/log 
GPU-ready log function
 2017-11-23 17:17:39 +00:00
Mike J Innes
9f5c4dd3e9
Merge pull request #104 from baggepinnen/patch-1 
Allow array of optimisers to train!
 2017-11-21 17:16:35 +01:00
Mike J Innes
351d3d4771 std derivative 2017-11-21 17:04:04 +01:00
Mike J Innes
b06884b912 LayerNorm tweaks 2017-11-21 16:32:36 +01:00
skariel
11d53781b2 adding layer normalization 2017-11-21 16:30:24 +01:00
Mike J Innes
979949d01a style 2017-11-21 15:25:09 +01:00
Fredrik Bagge Carlson
8991ce028c
Fix bug in rmsprop and adadelta 
`@. p.Δ = η * p.Δ / √acc` parses correctly while `@. p.Δ /= √acc*η` seems to parse like `@. p.Δ /= (√acc*η)`, hence the step size was de facto interpreted as `1/η`
 2017-11-14 17:32:16 +01:00
Mike J Innes
e0657d93ec mv numeric.jl to nnlib 2017-11-09 15:06:29 +00:00
Mike J Innes
2cb94981a0 gpu-ready log 2017-11-09 15:04:01 +00:00
Mike J Innes
e5d99d784e fixes #79 2017-11-09 14:53:26 +00:00
Mike J Innes
ccdc046546 fixes #79 2017-11-09 14:52:28 +00:00
Mike J Innes
752a9e2808 tree utilities 2017-11-08 22:19:01 +00:00
Mike J Innes
6eb2ec154b sentiment treebank loader 2017-11-08 22:19:01 +00:00
Mike J Innes
8777362eee exports 2017-11-08 22:19:01 +00:00
Mike J Innes
8b05317895 basic tree 2017-11-08 22:19:01 +00:00
Mike J Innes
7e9468d8f8 treebank skeleton 2017-11-08 22:19:01 +00:00
Mike J Innes
fcd091e8f0 Ac_mul_B derivatives 2017-11-08 22:18:45 +00:00
Mike J Innes
d4229c4815 useful params method 2017-11-08 22:18:45 +00:00
Mike J Innes
d6423eefe5 matrix-vector fast path 2017-11-08 22:18:45 +00:00
Fredrik Bagge Carlson
97244e0a68
Allow array of optimisers to train! 
This allows an array of optimisers to be sent to `train!`
 2017-11-04 13:27:32 +01:00
Mike J Innes
efa51f02e7 basic batch type 2017-11-02 11:49:42 +00:00
Mike J Innes
21ea93ffcd rename treelike 2017-11-02 11:47:34 +00:00
Iblis Lin
6c7613e02b batchnorm: leverage TrackedArray mean 2017-11-02 14:20:34 +08:00
Iblis Lin
88bd8a8fbd batchnorm: make CuArrays happy 2017-11-02 14:02:41 +08:00
Iblis Lin
477da75428 batchnorm: fix mapchildren 2017-11-02 13:32:12 +08:00
Iblis Lin
5253841acc batchnorm: update docs 2017-11-02 13:32:12 +08:00
Iblis Lin
b3356cc6bb batchnorm: batch σ correct coefficient 2017-11-02 13:32:12 +08:00
Iblis Lin
e0201be770 batchnorm: parameterize momentum and epsilon 2017-11-02 13:32:12 +08:00
Iblis Lin
669273b008 layer: implement BatchNorm layer 
See [Batch Normalization: Accelerating Deep Network Training by Reducing
Internal Covariate Shift](https://arxiv.org/pdf/1502.03167.pdf)
 2017-11-02 13:32:12 +08:00
Mike J Innes
e7a510da9a add cmudict dataset 2017-11-01 16:01:55 +00:00
Mike J Innes
0f8ba87dc6 treelike tuples 2017-10-31 16:37:41 +00:00
Mike J Innes
e943a39ee7 combine special cases 2017-10-31 16:37:33 +00:00
Iblis Lin
3d8b7250ae add scalar mean 2017-10-31 10:42:32 +00:00
Iblis Lin
c43bda019b TrackedArray: implement mean 
```julia
julia> p
Tracked 2×3 Array{Float64,2}:
 1.0  3.0  5.0
 2.0  4.0  6.0
```

Before
```julia
julia> @benchmark Flux.Tracker.back!(sum($p, 2) ./ size($p, 2), ones(2, 1))
BenchmarkTools.Trial:
  memory estimate:  3.44 KiB
  allocs estimate:  75
  --------------
  minimum time:     20.438 μs (0.00% GC)
  median time:      21.239 μs (0.00% GC)
  mean time:        22.354 μs (1.68% GC)
  maximum time:     3.811 ms (98.51% GC)
  --------------
  samples:          10000
  evals/sample:     1
```

After
```julia
julia> @benchmark Flux.Tracker.back!(mean($p, 2), ones(2, 1))
BenchmarkTools.Trial:
  memory estimate:  1008 bytes
  allocs estimate:  21
  --------------
  minimum time:     5.973 μs (0.00% GC)
  median time:      6.310 μs (0.00% GC)
  mean time:        6.630 μs (1.96% GC)
  maximum time:     680.709 μs (97.28% GC)
  --------------
  samples:          10000
  evals/sample:     6
```
 2017-10-30 16:21:02 +08:00
Mike J Innes
4c1b1eb18c Merge pull request #92 from CarloLucibello/drop 
add Dropout layer
 2017-10-26 12:07:28 +01:00
Mike J Innes
84efbbcc84 tracker predicate tweaks 2017-10-26 12:06:29 +01:00
Mike J Innes
cf6b930f63 reorganise 2017-10-26 11:46:12 +01:00
Mike J Innes
0df300299f clearer error message, fixes #93 2017-10-26 11:15:14 +01:00
GenaBitu
df06c3351d
Merge branch 'master' into cat-fix 2017-10-26 00:52:29 +02:00
CarloLucibello
711ea09d99 address comments 2017-10-25 02:35:27 +02:00
CarloLucibello
536ab3861d setmode! -> testmode! 2017-10-23 16:23:29 +02:00
CarloLucibello
00a9e5f01f construct TrackedScalar with params(1) 2017-10-23 10:49:45 +01:00
CarloLucibello
86c7c9246e add == and < for tracked arrays 2017-10-23 11:41:08 +02:00
CarloLucibello
2e1ed4c3fc add dropout 2017-10-23 10:12:53 +02:00
Mike J Innes
2a66545ef8 rnn state reset 2017-10-19 17:21:08 +01:00
Mike J Innes
99a7697d13 adam eta default arg 2017-10-19 14:31:34 +01:00
Mike J Innes
e5c8f6d835 only export known good optimisers 2017-10-19 11:26:11 +01:00
Mike J Innes
5b6a5667ed tracked array restructure 2017-10-18 22:54:58 +01:00
Mike J Innes
c8d4844da4 chunk util 2017-10-18 17:07:58 +01:00
Mike J Innes
07ad7cfa40 learning rate as default arg 2017-10-18 17:07:49 +01:00
Mike J Innes
e82428bb83 batching docs 2017-10-18 16:40:14 +01:00
Mike J Innes
b817ce632c syntax highlighting 2017-10-18 15:44:06 +01:00
Mike J Innes
fd249b773e rnn docs 2017-10-18 15:30:05 +01:00
Mike J Innes
190f48a709 nnlib docs 2017-10-18 14:40:58 +01:00
Mike J Innes
12944ae125 nnlib exports 2017-10-18 12:56:58 +01:00
Mike J Innes
0fbc8dff61 typoe 2017-10-18 12:48:58 +01:00
Mike J Innes
d6dd27dae5 dense layer example 2017-10-18 12:47:45 +01:00
Mike J Innes
7426faf37d optimiser docs 2017-10-18 12:09:48 +01:00
CarloLucibello
041079237e add docsting to train! 2017-10-17 21:04:18 +01:00
CarloLucibello
6d3a2a2210 change argument name for better clarity 2017-10-17 21:04:18 +01:00
Mike J Innes
23674b2555 logitcrossentropy tweaks 2017-10-17 17:58:32 +01:00
pevnak
4aa7741ba9 logit cross entropy 2017-10-17 17:57:46 +01:00
Mike J Innes
6dff8ca8d3 rename crossentropy loss 2017-10-17 17:36:18 +01:00
Mike J Innes
1800c8f523 deprecate mapparams 2017-10-17 17:35:30 +01:00
Mike J Innes
949fd9ba97 loss function tweaks 2017-10-17 17:30:11 +01:00
Mike J Innes
c764b74eba rename and fix mapleaves 2017-10-17 01:08:15 +01:00
Mike J Innes
7aa0b43ceb onehot sanity check 2017-10-17 00:07:58 +01:00
Mike J Innes
e02e320008 more general fmap 2017-10-17 00:07:15 +01:00
Mike J Innes
64e242e96c export param 2017-10-16 08:53:46 +01:00
Mike J Innes
d3db051ca0 flip 2017-10-16 08:53:39 +01:00
Mike J Innes
9a155abecd batch and batchseq apis 2017-10-15 23:44:40 +01:00
Mike J Innes
646720cd05 fix 2017-10-15 23:44:16 +01:00
Mike J Innes
c6556a29e6 order-stable params 2017-10-10 12:16:32 +01:00
GenaBitu
ef6d10886d Exposed all optimisers 2017-10-06 14:20:09 +01:00
GenaBitu
2084df96ae
Merge branch 'master' into cat-fix 2017-10-06 15:00:26 +02:00
pevnak
bfcc1ac25d exposing optimisers 2017-10-05 12:36:18 +01:00
Mike J Innes
1abc4febe6 more general adaptors 2017-10-04 18:55:56 +01:00
Dave Kleinschmidt
2b95aff158 actually use init argument in LSTMCell 2017-10-03 19:26:42 +01:00
Mike J Innes
5fd1b7d9a2 remove gc hack 2017-10-02 20:50:18 +01:00
Mike J Innes
1b91e6b38d store onehotmatrix height 2017-10-02 20:50:11 +01:00
Mike J Innes
7c8dba0b85 gc in training loop 2017-09-27 23:14:58 +01:00
Mike J Innes
a32ae4914c onehotmatrix cuda support 2017-09-27 22:51:00 +01:00
Mike J Innes
a60a754d68 beginnings of gpu support 2017-09-27 21:58:34 +01:00
Mike J Innes
120a6db2bb Merge branch 'master' of github.com:MikeInnes/Flux.jl 2017-09-27 21:16:23 +01:00
Mike J Innes
4bafa2b374 generic tree functions 2017-09-27 21:11:21 +01:00
Mike J Innes
2ec8401d2c remove compiler 2017-09-27 20:48:39 +01:00
Mike J Innes
94e38c05b8 more informative 2017-09-27 18:33:23 +01:00
GenaBitu
136f9bbf74
Hack which doesn't break backprop 2017-09-22 11:47:04 +02:00
GenaBitu
a5fe5b6e65
Added multi-variable vcat for TrackedVector 2017-09-22 11:22:21 +02:00
Mike J Innes
f2052739c1 tweaks 2017-09-12 14:11:03 +01:00
Mike J Innes
a3fe89e348 rnn tanh by default 2017-09-12 13:12:25 +01:00
Mike J Innes
6728295355 Merge pull request #63 from JobJob/rnncell-args 
Enables passing an activation function to RNN/RNNCell
 2017-09-12 13:10:43 +01:00
Mike J Innes
28bbef81b9 f 2017-09-12 13:06:32 +01:00
Mike J Innes
972ecab9f9 rm Over Seq 2017-09-12 13:03:16 +01:00
Joel Mason
00439555d1 Enables passing an activation function to RNN/RNNCell 
Also, fixes it not using the `init` function provided
 2017-09-12 20:54:56 +10:00
Mike J Innes
b9652f1812 typo, fixes #61 2017-09-12 10:45:07 +01:00
Mike J Innes
5f24d61ba3 important 2017-09-11 14:10:12 +01:00
Mike J Innes
7041ab9960 rm chainseq 2017-09-11 14:02:43 +01:00
Mike J Innes
c80fb999ff one hot docs 2017-09-11 13:40:11 +01:00
Mike J Innes
3f83be7bb7 more flexible training loop 2017-09-11 13:11:55 +01:00
Mike J Innes
33a5d26e57 chain utility note 2017-09-09 20:02:48 -04:00
Mike J Innes
fedee95b14 docs updates 2017-09-09 19:58:32 -04:00
Mike J Innes
a36d6d2af3 layer docs 2017-09-08 17:52:41 -04:00
Mike J Innes
f55b8cd20e track -> param 2017-09-07 15:13:04 -04:00
Mike J Innes
085d3aa9b4 handle epoch elsewhere 2017-09-07 00:29:55 -04:00
Mike J Innes
aeaa138b6d cb convenience 2017-09-07 00:27:16 -04:00
Mike J Innes
e837bb0745 rnn stuff 2017-09-07 00:27:04 -04:00
Mike J Innes
a93c440c1e style 2017-09-06 23:09:39 -04:00
Mike J Innes
cca4d25a10 efficient traversal 2017-09-06 23:09:32 -04:00
Mike J Innes
7cfc42d166 grad refactor 2017-09-06 21:21:35 -04:00
Mike J Innes
3ef72a9d7b utils updates 2017-09-06 18:59:43 -04:00
Mike J Innes
4083c34547 seq stuff 2017-09-06 18:59:36 -04:00
Mike J Innes
1855a37319 onehot 2017-09-06 18:58:55 -04:00
Mike J Innes
d7e3f7d6e1 fix stack/squeeze usage 2017-09-06 16:02:38 -04:00
Mike J Innes
1946c46e29 basic seq functionality 2017-09-06 14:03:25 -04:00
Mike J Innes
2c8b7bc64b remove these for now 2017-09-06 14:03:12 -04:00
Mike J Innes
4c12d18033 yet another vcat method 2017-09-05 19:25:42 -04:00
Mike J Innes
c95e9376a5 constructors 2017-09-05 19:25:34 -04:00
Mike J Innes
b023da1b7d lstm initialisation 2017-09-05 02:42:32 -04:00
Mike J Innes
61de692b50 lstm nonzero hidden state 2017-09-05 02:37:48 -04:00
Mike J Innes
ec02f1fabd batching in rnns 2017-09-05 02:29:31 -04:00
Mike J Innes
830d7fa611 vcat fix 2017-09-05 02:28:11 -04:00
Mike J Innes
363caeddc6 repmat forward 2017-09-05 02:12:53 -04:00
Mike J Innes
a322c07fc8 vcat back 2017-09-05 02:11:28 -04:00
Mike J Innes
788d7d35f0 better numeric grads 2017-09-03 17:10:35 -04:00
Mike J Innes
8f4ccdd5ba scalar getindex backprop 2017-09-03 17:10:23 -04:00
Mike J Innes
47ba702747 tweak optimiser interface 2017-09-03 17:10:04 -04:00
Mike J Innes
d4211b1f23 sgd export 2017-09-03 17:09:53 -04:00
Mike J Innes
f33a8edd25 meh 2017-09-03 02:45:46 -04:00
Mike J Innes
e57ae77bbb juno progress 2017-09-03 02:44:32 -04:00
Mike J Innes
bd5822fd71 cleaner lstm 2017-09-03 02:24:47 -04:00
Mike J Innes
cf58748680 nicer trackedarray type printing 2017-09-03 02:12:54 -04:00
Mike J Innes
9642ae8cd6 basic recurrence 2017-09-03 02:12:44 -04:00
Mike J Innes
f6771b98cd clearer name for dense 2017-09-02 16:50:11 -04:00
Mike J Innes
fe2b35facc add callbacks back 2017-09-01 23:59:44 -04:00
Mike J Innes
bf098d551c fuck 2017-09-01 23:41:44 -04:00
Mike J Innes
107d9daa8f add some non-differentiable functions 2017-09-01 23:33:05 -04:00
Mike J Innes
387686eb41 optimisers rework 2017-09-01 17:06:51 -04:00
Mike J Innes
892a779ed1 tracked transpose 2017-09-01 11:42:18 -04:00
Mike J Innes
b95dae1868 opt refactor 2017-08-31 14:55:23 -04:00
Mike J Innes
7cd13789dd fix removed import 2017-08-29 17:14:01 -04:00
ylxdzsw
97ecb26003 wip optimisers 2017-08-29 17:00:24 -04:00
Mike J Innes
7bba38274b Merge branch 'master' of github.com:MikeInnes/Flux.jl 2017-08-28 01:41:11 +01:00
Mike J Innes
0b89e1374c gpu-friendly 2017-08-28 01:40:59 +01:00
Mike J Innes
73166c52a0 cleaner broadcasting fix 2017-08-27 09:49:42 +01:00
Mike J Innes
12dc6b66c5 whoops 2017-08-24 22:23:05 +01:00
Mike J Innes
52f5f4a4c0 initial cuarrays integration 2017-08-24 17:00:48 +01:00
Mike J Innes
e7f26370d7 training tweaks 2017-08-24 16:10:04 +01:00
Mike J Innes
1526b13691 basic training loop 2017-08-24 11:42:29 +01:00
Mike J Innes
9ce0439943 better mse 2017-08-24 11:40:51 +01:00
Mike J Innes
23690e0083 not useful enough 2017-08-24 11:40:19 +01:00
Mike J Innes
d162e028bb utility method 2017-08-23 17:50:49 +01:00
Mike J Innes
e4e9794f5e loss function gradients 2017-08-23 17:50:43 +01:00
Mike J Innes
60c3090981 broadcasting fix 2017-08-23 17:21:02 +01:00
Mike J Innes
23c5a1b163 softmax gradient 2017-08-23 02:03:17 +01:00
Mike J Innes
5eee653a64 gradient checks 2017-08-23 01:43:45 +01:00
Mike J Innes
56ed6f5680 de-broadcasting 2017-08-23 00:25:19 +01:00
Mike J Innes
bafecfede1 sgd 2017-08-22 22:25:18 +01:00
Mike J Innes
f2dd7b0e90 fix include case 2017-08-22 17:18:27 +01:00
Mike J Innes
0ce8c0cee4 param collection 2017-08-22 17:13:03 +01:00
Mike J Innes
1179269355 remove old params 2017-08-22 15:21:08 +01:00
Mike J Innes
f43eef8d7d consistent case 2017-08-22 15:13:24 +01:00
Mike J Innes
cd45df1eca vector sum 2017-08-22 15:12:12 +01:00
Mike J Innes
97af9db181 scalar sum 2017-08-22 12:24:08 +01:00
Mike J Innes
ef681f16ea use nnlib for activations 2017-08-21 17:53:04 +01:00
Mike J Innes
3e0f45046c nicer show 2017-08-21 17:20:09 +01:00
Mike J Innes
227e41c37b gpu-friendly dualify 2017-08-21 16:35:39 +01:00
Mike J Innes
65a49188e6 mat-vec mul 2017-08-20 13:48:43 +01:00
Mike J Innes
09974caba0 extend affine with activation 2017-08-20 13:35:35 +01:00
Mike J Innes
8e59160df6 inferable broadcast 2017-08-20 13:35:20 +01:00
Mike J Innes
18e69b33c9 forwarddiff does these 2017-08-19 22:05:50 +01:00
Mike J Innes
ad0e0ea5a7 explicitly broadcast sigmoid 2017-08-19 22:04:47 +01:00
Mike J Innes
4a9dc40e7c simplify organisation 2017-08-19 20:52:29 +01:00
Mike J Innes
318e503d9b remove rnns 2017-08-19 20:52:17 +01:00
Mike J Innes
a581856954 remove params from compiler 2017-08-19 20:38:45 +01:00
Mike J Innes
b18d8cd08f rename chain.jl 2017-08-19 20:26:07 +01:00
Mike J Innes
0222103c7f fix these later 2017-08-19 20:05:22 +01:00
Mike J Innes
bd6bffde48 silo the compiler 2017-08-19 20:04:21 +01:00
Mike J Innes
8ed4d569b3 export from flux 2017-08-19 18:23:21 +01:00
Mike J Innes
f12b1d0ca1 avoid back! fallback 2017-08-19 17:40:07 +01:00
Mike J Innes
c9eb58f146 printing tweaks 2017-08-19 16:38:40 +01:00
Mike J Innes
de76472d99 exports 2017-08-19 16:20:53 +01:00
Mike J Innes
6c1a38e157 some more derivatives 2017-08-19 16:02:19 +01:00
Mike J Innes
380d32dac9 remove Core for now 2017-08-19 16:01:32 +01:00
Mike J Innes
70393138bc submodule 2017-08-19 11:11:25 +01:00
Mike J Innes
1889ccd316 tracker submodule 2017-08-19 11:06:19 +01:00
Mike J Innes
15c4f38130 printing 2017-08-19 11:00:55 +01:00
Mike J Innes
f1dda12a54 matmul 2017-08-19 10:14:50 +01:00
Mike J Innes
d9c30db2e3 AD skeleton 2017-08-18 16:50:27 +01:00
Mike J Innes
f8482ff80c remove back!, update! 2017-08-18 10:18:45 +01:00
Mike J Innes
21089fea9c remove some explicit back! 2017-08-18 01:19:10 +01:00
Mike J Innes
052cc52ada remove inference, for now 2017-08-18 01:19:10 +01:00
Mike J Innes
e79a1657d4 remove batching and training 2017-08-18 01:19:06 +01:00
Mike J Innes
5f9d8702a4 remove old cb macro 2017-08-18 01:07:58 +01:00
Mike J Innes
bfe1f8eeee shouldn't do this 2017-08-18 01:05:46 +01:00
Mike J Innes
4ed0929ac1 to be rewritten 2017-08-18 01:05:46 +01:00
Mike J Innes
8869f1f953 these are now useless 2017-08-18 01:05:46 +01:00
Mike J Innes
536949891d remove backends 2017-08-18 01:05:42 +01:00
Mike J Innes
cd9521a762 Merge branch 'master' into tf-train 2017-08-17 23:52:12 +01:00
Mike J Innes
f5b41d2e89 Merge pull request #50 from alha02/add-more-tf-ops 
Add more tf ops
 2017-08-17 21:47:47 +01:00
ylxdzsw
74a4a48162 fix 2017-08-01 13:28:14 +08:00
Mike J Innes
3000c7bbcb style tweak 2017-07-27 20:43:53 +01:00
ylxdzsw
88fa163c95 throttle 2017-07-21 16:31:12 +08:00
ylxdzsw
dd82edd818 fix training on the tensorflow backend 2017-07-21 11:22:40 +08:00
Ali Hamdi
ec7a0bd8f7 add expand_dims 2017-07-08 11:59:17 +02:00
Ali Hamdi
6d106c914d add randn 2017-07-08 11:59:17 +02:00
Ali Hamdi
cd8c2c2f83 add randu 2017-07-08 11:59:17 +02:00
Ali Hamdi
c8d320233a add triangular_solve 2017-07-08 11:59:17 +02:00
Ali Hamdi
015e373a34 add solve 2017-07-08 11:59:17 +02:00
Ali Hamdi
edbf698ed9 add cast 2017-07-08 11:59:17 +02:00
Ali Hamdi
f25f985109 add fill 2017-07-08 11:59:17 +02:00
Ali Hamdi
2fce3b195e add tile 2017-07-08 11:59:17 +02:00
Ali Hamdi
7aad224206 add reshape 2017-07-08 11:58:31 +02:00
Mike J Innes
e915f614aa remove confusing comments 2017-07-03 20:47:41 +01:00
Mike J Innes
e3b432165c Merge pull request #43 from alha02/add-more-tf-ops 
Add tf ops: diag, diagm, size, chol
 2017-07-03 18:48:32 +01:00
Mike J Innes
7e4801832b Merge pull request #44 from ylxdzsw/train-naive 
training julia models
 2017-07-03 18:41:39 +01:00
Mike J Innes
ea5d43ed77 show methods 2017-06-19 16:49:29 -07:00
Mike J Innes
0a6b6e1d6f cleaner chunk util 2017-06-19 16:49:17 -07:00
Mike J Innes
b3ef370456 inner constructor depwarn 2017-06-19 16:49:04 -07:00
Mike J Innes
5ee6e8f504 overload correct repeated 2017-06-19 16:48:52 -07:00
Mike J Innes
7d2a34b55d unroll multiple inputs 2017-06-17 19:21:39 -07:00
Mike J Innes
8306ed2ed7 don't force params here 2017-06-17 19:18:09 -07:00
ylxdzsw
c9ae219613 simplify back! of Chain 2017-06-14 21:58:43 +08:00
ylxdzsw
cca21a617c training julia models 2017-06-12 19:39:34 +08:00
Ali Hamdi
422b3bd5d8 Merge remote-tracking branch 'origin/add-more-tf-ops-2' into add-more-tf-ops 2017-06-10 11:45:30 +02:00
Ali Hamdi
5143410313 Merge remote-tracking branch 'upstream/master' into add-more-tf-ops-2 2017-06-10 11:43:27 +02:00
Mike J Innes
e4a7e7b335 more general seq/tuple support 2017-06-10 08:33:17 +01:00
Ali Hamdi
0b9436cfcd add chol 2017-06-09 21:50:25 +02:00
Ali Hamdi
7559196ab9 add size 2017-06-09 21:46:29 +02:00
Ali Hamdi
697a300ec3 add diagm 2017-06-09 21:35:31 +02:00
Ali Hamdi
6dfb554e95 add diag 2017-06-09 21:22:02 +02:00
Ali Hamdi
e6db3b0e89 Merge remote-tracking branch 'upstream/master' into add-more-tf-ops-2 2017-06-09 20:53:17 +02:00
Mike J Innes
358ba650ad more robust batches 2017-06-09 18:57:18 +01:00
Mike J Innes
65400f20ab nested batch tweaks 2017-06-09 18:55:21 +01:00
Mike J Innes
9c8dbb6b4b feedforward fix 2017-06-09 18:54:35 +01:00
Ali Hamdi
d80c1abf5e Merge remote-tracking branch 'upstream/master' into add-more-tf-ops-2 2017-06-09 17:55:00 +02:00
Mike J Innes
9b18fd639a mxarray ctx 2017-06-09 14:29:47 +01:00
Ali Hamdi
6b4db60f05 add length 2017-06-09 12:13:30 +02:00
Ali Hamdi
74b36b0094 add cat 2017-06-09 12:09:20 +02:00
Ali Hamdi
57c733af67 add permutedims 2017-06-09 11:55:26 +02:00
Ali Hamdi
d78963ab27 add transpose 2017-06-09 07:38:07 +02:00
Ali Hamdi
a098158ff5 Merge branch 'master' of https://github.com/MikeInnes/Flux.jl into add-more-tf-ops 2017-06-09 07:15:20 +02:00
Mike J Innes
1cc8100456 ctx methods for seq models 2017-06-09 01:35:07 +01:00
Mike J Innes
020ae616cc custom mxnet context 2017-06-09 00:56:52 +01:00
Mike J Innes
fe0bddd98d pass args correctly 2017-06-09 00:55:54 +01:00
Ali Hamdi
c350bfb672 Merge remote-tracking branch 'upstream/master' into add-more-tf-ops 2017-06-08 11:59:59 +02:00
Mike J Innes
4ccbbbb284 dup params fix 2017-06-08 10:49:39 +01:00
Mike J Innes
ab0f184d29 old import 2017-06-08 10:47:58 +01:00
Ali Hamdi
547ca64206 add det 2017-06-08 08:16:12 +02:00
Ali Hamdi
92616a8c3d add inv 2017-06-08 07:31:23 +02:00
Ali Hamdi
ee2e388228 add svd 2017-06-08 07:29:17 +02:00
Mike J Innes
dac80081ba these are probably the wrong abstraction 2017-06-06 18:03:36 +01:00
Mike J Innes
72e8e5c7f9 remove iterators 2017-06-06 18:03:14 +01:00
Mike J Innes
66c0403ffd seq rpad 2017-06-06 17:15:29 +01:00
Mike J Innes
e5df1c31dd batch vcat 2017-06-06 17:15:22 +01:00
Mike J Innes
3c52622896 more constructors 2017-06-06 16:58:57 +01:00
Mike J Innes
5186aa392f refactor 2017-06-06 15:59:10 +01:00
Mike J Innes
21fd79da16 fix 2017-06-05 22:51:08 +01:00
Mike J Innes
13f4af2492 generic and consistent conversions 2017-06-05 22:49:31 +01:00
Mike J Innes
bf3db99bc1 shared conversions 2017-06-05 19:04:14 +01:00
Mike J Innes
b2baf512e1 more code sharing 2017-06-05 18:46:36 +01:00
Mike J Innes
7505ee3a16 nicer display, share more code 2017-06-05 18:10:43 +01:00
Mike J Innes
ce6c96c2be rename 2017-06-05 18:00:44 +01:00
Mike J Innes
8e019e0f30 move batch and seq together 2017-06-05 17:51:30 +01:00
Mike J Innes
a7a46e3a82 that was what I wanted 2017-06-05 17:50:00 +01:00
Mike J Innes
957e23eb6b abstract batch type 2017-06-05 17:45:11 +01:00
Mike J Innes
da1697a257 fix juno import 2017-06-05 17:31:43 +01:00
Mike J Innes
1d7ac3ef3c tweak descriptions 2017-06-05 17:16:38 +01:00
Mike J Innes
2717ace397 split out core 2017-06-05 17:16:28 +01:00
Mike J Innes
9a460e12f2 make Batches submodule 2017-06-05 16:56:44 +01:00
Mike J Innes
cf8227c02f move @ Chain 2017-06-05 16:56:24 +01:00
Mike J Innes
099b9fbeb7 collect exports 2017-06-05 16:37:06 +01:00
Mike J Innes
415c5f6963 fix backend imports 2017-06-05 16:32:16 +01:00
Mike J Innes
ddcd576a74 give up and use AbstractArray 2017-06-05 16:09:06 +01:00
Mike J Innes
837173d65b model abstract is pretty useless 2017-06-05 16:08:23 +01:00
Mike J Innes
4685d2e672 strip down non-obvious exports 2017-06-05 15:47:26 +01:00
Mike J Innes
eafba7a55f graph for seqmodel 2017-06-03 15:52:43 +01:00
Mike J Innes
10abb64f4b fix stateful in backends 2017-06-02 16:28:31 +01:00
Mike J Innes
5232a3300c don't return hidden states 2017-06-02 16:22:55 +01:00
Mike J Innes
baaf9f2682 unrolled graphs only have one output 2017-06-02 16:22:39 +01:00
Mike J Innes
50bbb87674 stateless 2017-06-02 16:11:09 +01:00
Mike J Innes
d1f370a2f1 move stateful 2017-06-02 16:02:47 +01:00
Mike J Innes
2b85c76785 preserve original param states 2017-06-02 15:46:24 +01:00
Mike J Innes
af65b9200c mxnet repeated 2017-06-02 15:34:30 +01:00
Mike J Innes
4698cc0156 this might be a bit dirty 2017-06-02 15:22:23 +01:00
Mike J Innes
215e997540 broadcastto 2017-06-02 15:02:30 +01:00
Mike J Innes
edb1099cec more mxnet ops 2017-06-02 14:42:15 +01:00
Mike J Innes
e076bee00a mxnet broadcastto 2017-06-02 13:44:29 +01:00
Mike J Innes
4712568ac2 group tuple functions 2017-06-02 13:43:58 +01:00
Mike J Innes
4245f860e0 mxnet exp. 2017-06-02 13:42:59 +01:00
Mike J Innes
011449aa24 tuple sum 2017-06-02 13:40:38 +01:00
Mike J Innes
07c523a13d fix flatten/softmax batching behaviour 2017-06-01 19:28:02 +01:00
Mike J Innes
b54281bdea hadamard product 2017-06-01 19:27:46 +01:00
Mike J Innes
8cf5afdd79 add note 2017-06-01 18:37:20 +01:00
Mike J Innes
c7f8d86f9e fix mxnet 2017-06-01 16:57:39 +01:00
Mike J Innes
7a2a72a74a fix tensorflow 2017-06-01 16:56:48 +01:00
Mike J Innes
f7eb5179b1 fix basic interpreters 2017-06-01 16:55:48 +01:00
Mike J Innes
3532c7174f early throw on cyclic graphs 2017-06-01 16:52:49 +01:00
Mike J Innes
45c5502f90 obviate mapconst 2017-06-01 16:52:49 +01:00
Mike J Innes
c969d13306 mxnet vec 2017-05-31 18:23:33 +01:00
Mike J Innes
673abbcc1d mxnet hcat 2017-05-31 18:08:11 +01:00
Mike J Innes
033ed13f69 Merge pull request #38 from staticfloat/sf/act_on_your_feelings_young_one 
Fix activation functions
 2017-05-31 14:22:52 +01:00
Mike J Innes
b3e9dfa264 fix exec check 2017-05-30 18:04:31 +01:00
Mike J Innes
5654e15167 this is gone 2017-05-30 17:47:21 +01:00
Mike J Innes
790a58be1d support maps over tuples 2017-05-30 17:23:34 +01:00
Mike J Innes
d788cc8c54 remove old imap 2017-05-30 16:59:12 +01:00
Mike J Innes
5fa1eb4034 allow non-param states 2017-05-30 16:53:21 +01:00
Mike J Innes
b145b46cbb fix for recurrent anon models 2017-05-30 16:37:44 +01:00
Elliot Saba
dfa4dcccb5 Fix activation functions 2017-05-25 18:12:58 -07:00
Mike J Innes
e37973c3d5 Merge pull request #35 from oxinabox/patch-1 
Replace deprecated `unpack` with `unstack`
 2017-05-25 15:48:14 +01:00
Mike J Innes
fb8872b0b9 Merge pull request #24 from tkelman/nopiracy 
Don't extend base functions on base types
 2017-05-25 15:47:24 +01:00
Mike J Innes
9909af72a4 updates for new broadcasting approach 2017-05-24 12:02:03 +01:00
Lyndon White
d0fd9cd622 Replace deprecated unpack with unstack 2017-05-24 10:35:17 +08:00
Tony Kelman
41ea071f3a Use a package-local squeeze function instead of extending Base 2017-05-22 04:08:46 -04:00
Tony Kelman
5cbb47a13d Don't extend base functions on base types 
better broadcast syntax
 2017-05-22 04:05:57 -04:00
Mike J Innes
0ed4e65d2f refactor 2017-05-11 19:04:08 +01:00
ylxdzsw
1bd0a43b7d batch iterator 2017-05-11 19:03:53 +01:00
ylxdzsw
019e34194d catmat fix 2017-05-11 15:49:13 +01:00
Ali Hamdi
92d39f9ff7 dim defaults to nothing for reduction ops 2017-05-09 16:50:32 +02:00
Ali Hamdi
d5b073db74 add more basic math ops to tf backend 2017-05-09 14:39:10 +02:00
Ali Hamdi
eda5c2c776 add .^ and reduction ops for tf backend 2017-05-09 01:29:15 +02:00
Mike J Innes
bd97c0f93b strip away params 2017-05-04 17:01:10 +01:00
Mike J Innes
5be9ce45d8 support constant arrays in MXNet 2017-05-04 15:09:18 +01:00
Mike J Innes
a2db4b5319 factor out common tests 2017-05-04 13:52:31 +01:00
Mike J Innes
70286c0bf3 src docs 2017-05-04 13:17:21 +01:00
Mike J Innes
e04dcbd460 reorganise recurrent stuff 2017-05-04 10:45:44 +01:00
Mike J Innes
c025cddc73 runmodel no longer needed 2017-05-04 10:32:53 +01:00
Mike J Innes
51bf904179 better return value 2017-05-03 19:13:33 +01:00
Mike J Innes
c9dcc815dc add this for convenience 2017-05-03 18:33:23 +01:00
Mike J Innes
8c10a95bf4 more appropriate naming 2017-05-03 18:33:14 +01:00
Mike J Innes
d7ff193ad6 basic back pass 2017-05-01 19:44:58 +01:00
Mike J Innes
af30ee78d2 remove old code 2017-05-01 19:44:26 +01:00
Mike J Innes
b35f50571c use params instead of vars 2017-05-01 18:27:52 +01:00
Mike J Innes
796d7d7e99 recurrent models in tf 2017-05-01 18:05:17 +01:00
Mike J Innes
3998be2244 tuple i/o tests 2017-05-01 17:44:20 +01:00
Mike J Innes
2467ca4187 tuple inputs in tensorflow 2017-05-01 17:41:42 +01:00
Mike J Innes
357f989de5 pull out tuple utils 2017-05-01 16:57:51 +01:00
Mike J Innes
2934607115 multi-tuple output 2017-05-01 16:30:24 +01:00
Mike J Innes
f5b0f3e530 identity function 2017-05-01 16:28:39 +01:00
Mike J Innes
ba54552be5 remove special-cased training code 
splits model and exec to allow multi inputs
 2017-05-01 16:12:46 +01:00
Mike J Innes
d4ee8a6a2a use juno's info function 2017-05-01 15:14:29 +01:00
Mike J Innes
b19e31714d allow access to loop vars 2017-05-01 14:00:39 +01:00
Mike J Innes
33c6781c99 remove from training.jl 2017-05-01 13:47:08 +01:00
Mike J Innes
6778d00dbf this is no longer specific to training 2017-05-01 13:46:23 +01:00
Mike J Innes
d095c640c9 this should be a general util 2017-05-01 13:45:24 +01:00
Mike J Innes
eae13c533f training callbacks 2017-05-01 13:43:07 +01:00
Mike J Innes
5dce8df678 fix accuracy for batches 2017-05-01 13:40:11 +01:00
Mike J Innes
dba6bef245 this should be progress'd too 2017-05-01 12:46:02 +01:00
Mike J Innes
38852964f6 organise training and utils 2017-05-01 12:41:54 +01:00
Mike J Innes
0e6bb17709 import this in compiler 2017-05-01 12:31:00 +01:00
Mike J Innes
36f4103d7d cost is really a layer 2017-05-01 12:30:28 +01:00
Mike J Innes
7425584c91 more correct 2017-05-01 11:33:24 +01:00
Mike J Innes
7415e61287 remove mse!, export logloss 2017-05-01 11:28:41 +01:00
Mike J Innes
360d128b1a logloss 2017-04-28 17:14:27 +01:00
Mike J Innes
ef4ec5be4b customisable loss 2017-04-28 17:14:21 +01:00
Mike J Innes
63b328142a print epochs again 2017-04-27 17:43:38 +01:00
Mike J Innes
121af0579a update for rnn wrappers 2017-04-27 17:27:08 +01:00
Mike J Innes
f4acd286c8 fix tensorflow dep 2017-04-27 12:48:11 +01:00
Mike J Innes
52a7199d10 gradients for recurrent models 2017-04-26 17:42:47 +01:00
Mike J Innes
19cf3e2b62 split out runseq 2017-04-19 17:33:55 +01:00
Mike J Innes
edfb0211e6 better for nested batches 2017-04-19 17:18:40 +01:00
Mike J Innes
14afe54143 fixes for recurrent networks 2017-04-19 17:17:37 +01:00
Mike J Innes
358334a893 mxnet back! for multi outputs 2017-04-19 17:13:57 +01:00
Mike J Innes
42a8117704 seqmodel with raw arrays 2017-04-19 15:13:17 +01:00
Mike J Innes
f8a3b02c1d fix tf train 2017-04-19 14:48:10 +01:00
Mike J Innes
42688f8aa8 update training process, mnist example 2017-04-19 14:23:48 +01:00
Mike J Innes
f6334e7e96 still need batchone 2017-04-19 13:26:37 +01:00
Mike J Innes
5357b1e9f9 remove fake batching vestiges 2017-04-19 13:19:18 +01:00
Mike J Innes
228f7d487c remove fake batch semantics 2017-04-18 21:04:21 +01:00
Mike J Innes
88cf6d9e61 sensible default for unsqueeze 2017-04-18 20:55:59 +01:00
Mike J Innes
63f3f3b569 no longer needed 2017-04-18 19:27:24 +01:00
Mike J Innes
a3acf11a60 remove ml export 2017-04-18 19:02:55 +01:00
Mike J Innes
5cb30cecfe nicer convertel 2017-04-17 18:15:01 +01:00
Mike J Innes
e39c24fd86 lazy loading note 2017-04-17 15:56:59 +01:00
Mike J Innes
c22561c7eb Merge pull request #15 from ranjanan/RA/mse 
Fix tiny bug in mse cost function
 2017-04-04 19:23:44 +01:00
Mike J Innes
efa8d75ea3 neater 2017-04-04 19:19:08 +01:00
Mike J Innes
f8e1f20728 mxnet recurrence test 2017-03-31 12:39:23 +01:00
Mike J Innes
b4221f6ea6 recurrence working 2017-03-30 20:05:18 +01:00
Mike J Innes
4113d4d476 mimo working in mxnet 2017-03-30 19:50:03 +01:00
Mike J Innes
94e384930d fix back pass 2017-03-30 19:36:59 +01:00
Mike J Innes
4de61fc377 fix exec forward pass 2017-03-30 19:25:54 +01:00
Mike J Innes
4df97bf607 build executor correctly 2017-03-30 19:16:24 +01:00
Mike J Innes
5df56b6073 build graphs lazily 2017-03-30 18:14:08 +01:00
Mike J Innes
acbc4ea071 mimo models 2017-03-30 15:54:42 +01:00
ranjanan
c773229136 Fix tiny bug in mse cost function 2017-03-30 12:09:40 +05:30
Mike J Innes
ca0e20ed7a make seqmodel a model 2017-03-29 19:25:50 +01:00
Mike J Innes
5111db4037 unrolling test 2017-03-29 18:30:28 +01:00
Mike J Innes
7e983c74cb make this immutable again 2017-03-29 17:47:43 +01:00
Mike J Innes
2082d9db5c working seqmodel 2017-03-28 19:54:32 +01:00
Mike J Innes
abe66d398f remove old recurrence code 2017-03-27 18:24:02 +01:00
Mike J Innes
05ac3855f4 fixes 2017-03-27 18:23:22 +01:00
Mike J Innes
1a5e050a88 recurrence overhaul mostly working 2017-03-21 01:32:12 +00:00
Mike J Innes
90edefe072 use broadcasting plus 2017-03-21 01:18:00 +00:00
Mike J Innes
a5bd72753e fix nested mapt 2017-03-20 23:19:03 +00:00
Mike J Innes
eadf96605c fix recurrent layer batching 2017-03-20 23:11:08 +00:00
Mike J Innes
6237aa6739 shape: handle tuples better 2017-03-20 23:10:38 +00:00
Mike J Innes
6756ce7528 relu inferrence 2017-03-20 19:57:11 +00:00
Mike J Innes
498a66e7b6 update capacitors 2017-03-20 19:57:00 +00:00
Mike J Innes
7af64398d5 split out unseqout 2017-03-19 22:12:30 +00:00
Mike J Innes
8f911cc31e so can unseq 2017-03-19 22:12:30 +00:00
Mike J Innes
69b24bfa9b stateless can be a postprocess 2017-03-19 22:12:30 +00:00
Mike J Innes
93ed44b31a note about unroll options 2017-03-19 22:12:29 +00:00
Mike J Innes
d73e962da9 replace old shape inference system 2017-03-17 16:34:54 +00:00
Mike J Innes
7ef6e295cb reuse mapt 2017-03-17 01:04:41 +00:00
Mike J Innes
d27d59b071 softmax on batches 2017-03-16 11:52:00 +00:00
Mike J Innes
2adc3cd18e new struct syntax 2017-03-14 17:56:03 +00:00
Mike J Innes
0cf99dbbdf 0.6 dep 2017-03-14 17:20:39 +00:00
Mike J Innes
b0ae0849d0 invoke dep 2017-03-14 16:52:49 +00:00
Mike J Innes
6751657dfa typealias / abstract deps 2017-03-14 16:51:31 +00:00
Mike J Innes
fa56f78781 isa syntax 2017-03-14 15:21:18 +00:00
Mike J Innes
daf962a831 tensorflow exception handling 2017-03-12 18:34:11 +00:00
Mike J Innes
4b82c57f88 strip more of the trace 2017-03-12 18:33:47 +00:00
Mike J Innes
cde74337c8 should've moved this 2017-03-12 18:33:29 +00:00
Mike J Innes
06e9626c0b strip macro line numbers 2017-03-12 18:33:02 +00:00
Mike J Innes
418f5bfe28 restore alterparam 2017-03-12 14:51:55 +00:00
Mike J Innes
3f919aa81e avoid line numbers in runmodel 2017-03-09 01:10:44 +00:00
Mike J Innes
15b3ce1ada factor out fake batching semantics 2017-03-09 00:13:26 +00:00
Mike J Innes
c4d815b5fc move todo 2017-03-09 00:12:49 +00:00
Mike J Innes
854a1e1886 nicer mxnet api 2017-03-08 21:41:13 +00:00
Mike J Innes
9d1d176749 transpose everything going into mxnet 2017-03-08 17:35:15 +00:00
Mike J Innes
3b004bac7d fix for latest mxnet api 2017-03-08 16:20:33 +00:00
Mike J Innes
6a3bed1e61 this is pretty useless 2017-03-08 15:38:55 +00:00
Mike J Innes
9f62227db3 mxarray setindex 2017-03-08 15:36:51 +00:00
Mike J Innes
d713c67f02 redundant method 2017-03-08 15:36:44 +00:00
Mike J Innes
39bb625246 better mxarray dims api 2017-03-08 15:36:38 +00:00
Mike J Innes
9ecb1311a5 copy methods for param 2017-03-08 15:36:25 +00:00
Mike J Innes
d35191595d mxarray 2017-03-08 01:19:51 +00:00
Mike J Innes
06f2ee2284 iterate over chain 2017-03-07 14:37:37 +00:00
Mike J Innes
2a57150bce AvgPool shim 2017-03-06 17:21:35 +00:00
Mike J Innes
5d919175fc mxnet multi output 2017-03-06 17:20:15 +00:00
Mike J Innes
d9910070b4 fix/test native multi outputs 2017-03-06 16:12:03 +00:00
Mike J Innes
a03898d24d this was wrong 2017-03-03 14:06:51 +00:00
Mike J Innes
4d4979b401 better alternative to basemodel 2017-02-28 16:42:48 +00:00
Mike J Innes
5f1f2ebaa2 model storage notes 2017-02-28 16:41:33 +00:00
Mike J Innes
4c0e915c00 remove c++ stacktrace 2017-02-27 19:49:32 +00:00
Mike J Innes
c33f695c51 precompile 2017-02-24 15:49:03 +00:00
Mike J Innes
f3d0c21952 quote was the right thing here 2017-02-24 15:48:52 +00:00
Mike J Innes
bad6b2d1ae compile the forward pass again 2017-02-24 14:38:17 +00:00
Mike J Innes
9921881d47 fixes apparently needed for 0.6 2017-02-23 23:40:22 +00:00
Mike J Innes
06fd5adddc parameter storage 2017-02-23 21:42:34 +00:00
Mike J Innes
a4812579e9 fix back pass 2017-02-23 21:06:46 +00:00
Mike J Innes
2d77220d60 tweak param loading 2017-02-23 18:48:46 +00:00
Mike J Innes
2f2ff0b03b graph struct 2017-02-23 17:32:06 +00:00
Mike J Innes
f230b7cebf make mxerr more robust 2017-02-23 16:58:29 +00:00
Mike J Innes
21775f8c89 fix for error message change 2017-02-23 16:58:17 +00:00
Mike J Innes
616425554d update for mxnet api 2017-02-23 16:58:10 +00:00
Mike J Innes
a794f068a5 try to get biases working somewhat naturally 2017-02-21 16:07:58 +00:00
Mike J Innes
092f2038b3 use affine only for compat 2017-02-21 14:12:11 +00:00
Mike J Innes
0bb44f5ace affine error messages 2017-02-21 13:31:41 +00:00
Mike J Innes
8d63bf8053 get this somewhat working 2017-02-21 12:58:31 +00:00
Mike J Innes
417a70713b julia's map is shitty 2017-02-21 09:05:08 +00:00
Mike J Innes
521258eb08 more intuitive naming 2017-02-21 08:56:05 +00:00
Mike J Innes
4fb51d1752 factor out var creation 2017-02-21 08:52:56 +00:00
Mike J Innes
12cde694b3 conv/affine hacks 2017-02-21 08:52:40 +00:00
Mike J Innes
96c0e76b92 style 2017-02-21 08:50:54 +00:00
Mike J Innes
14cbdc5e9c these aren't working 2017-02-21 08:50:47 +00:00
Mike J Innes
b49daa367f fix for empty dicts 2017-02-20 23:45:07 +00:00
Mike J Innes
54011045e7 fix 2017-02-20 23:15:27 +00:00
Mike J Innes
3fdffea37d fix 2017-02-20 21:50:01 +00:00
Mike J Innes
b92252cd5c more interp catches 2017-02-20 21:49:47 +00:00
Mike J Innes
242b4d6e23 fixes 2017-02-20 21:49:02 +00:00
Mike J Innes
ad4d60f90d fix mx.FeedForward 2017-02-20 19:35:32 +00:00
Mike J Innes
f3a9934858 update mnist example 2017-02-02 10:09:41 +05:30
Mike J Innes
a041da3fdc shapes macro 2017-02-01 21:06:02 +05:30
Mike J Innes
cd86dfdf07 fix hintify 2017-02-01 20:26:38 +05:30
Mike J Innes
a1e35ea2c8 stacktrace fixes 2017-02-01 19:51:08 +05:30
Mike J Innes
00a5acde46 fixes 2017-02-01 17:06:42 +05:30
Mike J Innes
7c450c3200 clean up forward pass interp errors 2017-02-01 16:56:53 +05:30
Mike J Innes
abe4ab94aa clean up interpreter imports 2017-02-01 12:27:02 +05:30
Mike J Innes
7750f5375b handle interpreter errors 2017-02-01 11:56:20 +05:30
Mike J Innes
1f3587e9dc mx batch semantics 2017-01-30 23:35:15 +05:30
Mike J Innes
16d6c9aed9 MXModel -> MX.Model 2017-01-30 23:35:05 +05:30
Mike J Innes
3cb3aea825 pretty sure we don't need this 2017-01-30 23:34:36 +05:30
Mike J Innes
cd0aa26b0e simplify graph 2017-01-30 23:19:18 +05:30
Mike J Innes
de72d83f7c factor out node registration 2017-01-30 23:12:01 +05:30
Mike J Innes
508cd730c9 use ituple properly 2017-01-30 22:55:12 +05:30
Mike J Innes
a73b53e05e use iargs 2017-01-30 22:51:49 +05:30
Mike J Innes
7a01605aab add this back later 2017-01-30 22:38:38 +05:30
Mike J Innes
5f1d560e38 265 2017-01-30 22:26:46 +05:30
Mike J Innes
349e81172d throw graph-aware exceptions 2017-01-29 23:35:03 +05:30
Mike J Innes
3981485500 show compile error trace 2017-01-29 17:29:37 +05:30
Mike J Innes
c2d6059d73 redundant 2017-01-29 16:58:22 +05:30
Mike J Innes
12d05a2db1 record stack traces 2017-01-29 16:09:30 +05:30
Mike J Innes
8bf5d91605 symbolname utility 2017-01-28 23:36:58 +05:30
Mike J Innes
70168319eb remove tond 2017-01-28 23:07:22 +05:30
Mike J Innes
bb70f401be remove old shape hacks 2017-01-28 23:07:02 +05:30
Mike J Innes
94cb98c13f basic mxnet backend 2017-01-28 22:32:49 +05:30
Mike J Innes
3b3a088851 more correct 2017-01-27 17:03:22 +05:30
Mike J Innes
a71c79e920 convert parameters also 2017-01-27 16:02:52 +05:30
Mike J Innes
42fabadd11 test native tensor support 2017-01-27 00:02:59 +05:30
Mike J Innes
b79e536c13 convertel for batch 2017-01-26 23:37:06 +05:30
Mike J Innes
8db503eafa make tf model running easier 2017-01-25 18:10:56 +05:30
Mike J Innes
4c8922270f todo 2017-01-25 15:33:58 +05:30
Mike J Innes
56c5784d83 batching for basic layers 2017-01-24 17:23:42 +05:30
Mike J Innes
568b8d7e48 organise batching utils 2017-01-24 15:54:30 +05:30
Mike J Innes
d7fe525f4d fix the build 2017-01-16 01:21:45 +01:00
Mike J Innes
525ebf00f2 make use of line info 2016-12-27 00:45:36 +00:00
Mike J Innes
8d1171cb96 shape debugger 2016-12-26 18:55:43 +00:00
Mike J Innes
87e928638a remove expand 2016-12-26 14:48:24 +00:00
Mike J Innes
147a26d045 interpreter middleware refactor 2016-12-26 13:42:12 +00:00
Mike J Innes
3bfb2c5a37 more stack frame integration 2016-12-26 12:11:24 +00:00
Mike J Innes
353f156354 split interpret / interpv 2016-12-21 13:05:18 +00:00
Mike J Innes
6acfcd913e fix self escaping 2016-12-21 13:04:54 +00:00
Mike J Innes
9775baddd3 print last line for TF errors 2016-12-20 18:15:42 +00:00
Mike J Innes
b580c2e4a7 style improvement 2016-12-20 17:33:14 +00:00
Mike J Innes
1b5b28897c store stacks as part of model 2016-12-20 17:32:33 +00:00
Mike J Innes
0e08f175bc makesession refactor 2016-12-20 17:18:40 +00:00
Mike J Innes
22568452f1 remove gradient support for now 2016-12-20 17:18:02 +00:00
Mike J Innes
f74ca7f7cf collect line info in tensorflow 2016-12-20 16:31:20 +00:00
Mike J Innes
5f27e30e68 basic line node handling 2016-12-20 15:44:00 +00:00
Mike J Innes
17449e15a3 macrotools update 2016-12-19 15:04:41 +00:00
Mike J Innes
bbee60cd86 onehot/cold docs 2016-12-16 15:54:55 +00:00
Mike J Innes
ac374087b3 load tf lazily to speed up load time 2016-12-16 15:54:15 +00:00
Mike J Innes
c82716b535 move activations 2016-12-15 23:24:10 +00:00
Mike J Innes
42ce2fadf1 don't do + twice 2016-12-15 23:08:56 +00:00
Mike J Innes
6114b70f76 use regular + 2016-12-15 22:57:36 +00:00
Mike J Innes
de6c3ef07e remove diff code for now 2016-12-15 22:39:48 +00:00
Mike J Innes
f3555a9c57 remove type rendering overload 2016-12-15 22:36:40 +00:00
Mike J Innes
1c21a860e2 some docstrings 2016-12-15 22:31:27 +00:00
Mike J Innes
62fd13bded consistently use delta for gradients 2016-12-15 21:37:39 +00:00
Mike J Innes
a330b394bd move batchone util 2016-12-15 21:37:07 +00:00
Mike J Innes
03840d043c fix ops 2016-12-15 20:53:15 +00:00
Mike J Innes
4b64bf11a5 fix lstm 2016-12-15 20:53:08 +00:00
Mike J Innes
c6fb9c1f0c fix model call 2016-12-15 18:35:11 +00:00
Mike J Innes
1b22d55401 fix param interpretation 2016-12-13 15:46:34 +00:00
Mike J Innes
2aa8dfc208 tweak constants approach 2016-11-17 11:28:24 +00:00
Mike J Innes
1424b75e78 extra ops 2016-11-17 11:28:15 +00:00
Mike J Innes
b2d9514a9a make astuple more general 2016-11-17 11:27:17 +00:00
Mike J Innes
3e42db0cdf make interpmap composable 2016-11-17 11:26:58 +00:00
Mike J Innes
a6169ec2d0 stateless model support 2016-11-15 23:54:08 +00:00
Mike J Innes
2d90d04789 explicit hidden state batching 2016-11-15 23:44:11 +00:00
Mike J Innes
3c068744d2 get rid of Group 2016-11-15 21:09:58 +00:00
Mike J Innes
d86225ca47 working map interp 2016-11-15 20:46:01 +00:00
Mike J Innes
46aa456bd1 more robust loop lift 2016-11-15 20:45:24 +00:00
Mike J Innes
27554494b7 initial abstract interpretation 2016-11-15 20:01:56 +00:00
Mike J Innes
806febdd37 unroll1 2016-11-15 20:01:35 +00:00
Mike J Innes
7f2ef71daf interpreter 2016-11-15 16:58:59 +00:00
Mike J Innes
f31b539566 make these fit with julia semantics 2016-11-15 16:40:17 +00:00
Mike J Innes
b7caaf4a65 offset escaping 2016-11-15 16:39:51 +00:00
Mike J Innes
8fceb62d33 unrolled models without state input 2016-11-15 13:28:14 +00:00
Mike J Innes
85335e5ed0 esc self param 2016-11-15 00:32:50 +00:00
Mike J Innes
9062792495 rename affine file 2016-11-15 00:09:53 +00:00
Mike J Innes
bdd05157e2 dense -> affine 2016-11-14 22:16:00 +00:00
Mike J Innes
6140448f17 fix 2016-11-14 22:02:20 +00:00
Mike J Innes
ca5058f555 (╯°□°）╯︵ ┻━┻ 2016-11-14 21:58:16 +00:00
Mike J Innes
199765354e fix a todo, houdini-style 2016-11-14 21:56:40 +00:00
Mike J Innes
c597d3a793 ml macro 2016-11-14 20:14:53 +00:00
Mike J Innes
91652e5b44 need this for models with params 2016-11-14 15:42:29 +00:00
Mike J Innes
5a32c72362 use interpreter for forward pass 2016-11-13 20:46:35 +00:00
Mike J Innes
c654fe403a move towards abstract interpreter model 2016-11-13 20:27:20 +00:00
Mike J Innes
6ac4dd8429 move op 2016-11-13 18:16:58 +00:00
Mike J Innes
e4a6ca5f9e remove custom show 2016-11-13 15:35:20 +00:00
Mike J Innes
dab491955f using using 2016-11-11 01:01:26 +00:00
Mike J Innes
2bfe5cce8e update for macrotools change 2016-11-11 01:01:19 +00:00
Mike J Innes
600b3fb54c move input logic to flow 2016-11-09 21:47:54 +00:00
Mike J Innes
ad6e6b4116 update recurrent usage 2016-11-08 18:08:13 +00:00
Mike J Innes
224d84691d appears to be working 2016-11-08 18:02:14 +00:00
Mike J Innes
21a3b95260 mostly recover old behaviour 2016-11-08 00:06:45 +00:00
Mike J Innes
42a7a6ebf6 delay -> offset 2016-11-07 19:44:51 +00:00
Mike J Innes
c3d32c395c new progress integration 2016-11-04 22:28:12 +00:00
Mike J Innes
d7d95feab8 actually get GRU working 2016-11-02 00:36:13 +00:00
Mike J Innes
85415d4244 throw GRU together 2016-11-01 14:42:41 +00:00
Mike J Innes
53ebb5051a Flow -> DataFlow 2016-10-31 12:38:18 +00:00
Mike J Innes
7cd94b4a5d well, that was easy 😎 2016-10-31 11:01:19 +00:00
Mike J Innes
4517e41226 sampling + tweaks 2016-10-30 16:07:29 +00:00
Mike J Innes
508364407e simplify recurrent layer 2016-10-30 16:07:18 +00:00
Mike J Innes
a6fe1f3810 use new session and store params back in the model 2016-10-30 15:08:50 +00:00
Mike J Innes
b443425c6d cross entropy loss, loss checks 2016-10-30 14:12:03 +00:00
Mike J Innes
3b70ea6a42 split out makesession 2016-10-30 12:29:00 +00:00
Mike J Innes
e433ffce8f split out makesession logic 2016-10-30 12:10:44 +00:00
Mike J Innes
81d9743836 export recurrent 2016-10-30 11:41:52 +00:00
Mike J Innes
fabff9b0d2 todo 2016-10-30 11:38:35 +00:00
Mike J Innes
671640acd6 ninputs utility 2016-10-30 11:38:13 +00:00
Mike J Innes
a99bb03830 gradients are slow 2016-10-30 10:55:07 +00:00
Mike J Innes
ec1950b466 .* in tf 2016-10-30 10:54:55 +00:00
Mike J Innes
a1b1d87767 update states references 2016-10-30 01:58:39 +01:00
Mike J Innes
4d45ee1bb9 length for iterators 2016-10-30 01:18:20 +01:00
Mike J Innes
1761e43bc4 handle state in training 2016-10-30 00:24:29 +01:00
Mike J Innes
605e3a9363 don't rebatch batches 2016-10-30 00:20:15 +01:00
Mike J Innes
06e37081d3 consistent naming 2016-10-30 00:19:57 +01:00
Mike J Innes
73ff5b4201 batched training for char-rnn 2016-10-29 23:36:39 +01:00
Mike J Innes
ba78b99b86 include that 2016-10-29 00:16:24 +01:00
Mike J Innes
89c4a6df31 this is no longer test code 2016-10-29 00:13:32 +01:00
Mike J Innes
4de16171db basic sequence model training 2016-10-29 00:10:27 +01:00
Mike J Innes
d9abb8f0ce chunks util 2016-10-28 21:47:57 +01:00
Mike J Innes
d9ed5676c2 handle state on julia side 2016-10-28 21:17:48 +01:00
Mike J Innes
e450a585b7 handling of multiple outputs 2016-10-28 20:50:27 +01:00
Mike J Innes
1c6eaece5d rename seqmodel 2016-10-28 19:11:38 +01:00
Mike J Innes
102e09d14b tf recurrent models 2016-10-28 17:14:57 +01:00
Mike J Innes
217e28653a tf cycle conversion error 2016-10-28 17:12:19 +01:00
Mike J Innes
c5a64391a1 use batching api 2016-10-28 17:00:31 +01:00
Mike J Innes
2852dddf0f put this back 2016-10-28 16:26:06 +01:00
Mike J Innes
8140c2312d todone 2016-10-28 16:25:59 +01:00
Mike J Innes
d6eacf3375 better handling for reused params 2016-10-28 16:06:56 +01:00
Mike J Innes
740d868ef9 tf.model refactor 2016-10-28 15:13:58 +01:00
Mike J Innes
27aa2bf8d4 graph support 2016-10-28 15:13:43 +01:00
Mike J Innes
d42130b8cd fix for matmul 2016-10-28 15:02:48 +01:00
Mike J Innes
1a726033f4 do this properly 2016-10-26 15:49:35 +01:00
Mike J Innes
d5d7242c53 export unroll 2016-10-26 15:37:30 +01:00
Mike J Innes
0ad569596b tf reorg 2016-10-26 14:25:10 +01:00
Mike J Innes
82d69757c7 BatchSeq convenience alias 2016-10-26 12:37:48 +01:00
Mike J Innes
823792bc19 unrolled type 2016-10-26 11:57:03 +01:00
Mike J Innes
2a58b23085 more interesting recurrent model 2016-10-26 11:34:17 +01:00
Mike J Innes
652c26728e better loop lifting semantics 2016-10-26 11:19:45 +01:00
Mike J Innes
42b50c976a fix unrolling 2016-10-26 00:49:32 +01:00
Mike J Innes
ba60c4596b graph op fixes 2016-10-26 00:39:16 +01:00
Mike J Innes
eb78f67a93 refactor input model 2016-10-25 23:10:35 +01:00
Mike J Innes
10761a4bee update for flow 2016-10-25 22:28:30 +01:00
Mike J Innes
91a62a04bc update for flow exports 2016-10-25 21:32:51 +01:00
Mike J Innes
14e4117837 basic unrolling 2016-10-25 21:10:04 +01:00
Mike J Innes
1fde7b4615 preserve default values for hidden states 2016-10-25 19:10:26 +01:00
Mike J Innes
dea85df8b7 use param object rather than named input 2016-10-25 17:57:20 +01:00
Mike J Innes
ee0c5ae14e remove mxnet for now 2016-10-25 17:37:37 +01:00
Mike J Innes
d442dd8c5b use Float32 here 2016-10-25 16:23:04 +01:00
Mike J Innes
a06145a145 use new batching approach in TensorFlow 2016-10-25 16:21:17 +01:00
Mike J Innes
46550e4863 suspiciously similar seq data structure 2016-10-25 14:10:49 +01:00
Mike J Innes
95b955246d export rawbatch 2016-10-25 14:10:32 +01:00
Mike J Innes
7438ee6108 move convert method 2016-10-25 14:10:27 +01:00
Mike J Innes
1847809e99 batching refactor, nested batches 2016-10-25 13:48:30 +01:00
Mike J Innes
183c3b0680 batch tweaks 2016-10-15 18:16:04 +01:00
Mike J Innes
6d53b7af47 batch data structure 2016-10-12 22:49:08 +01:00
Mike J Innes
c9f9665e4e move batching logic 2016-10-12 17:07:22 +01:00
Mike J Innes
b115d8ce3f model -> net 2016-10-12 16:28:16 +01:00
Mike J Innes
a56af5d16e reshape layer 2016-10-10 23:48:16 +01:00
Mike J Innes
438dc9d40a fix conv2d shape inference 2016-10-10 23:20:40 +01:00
Mike J Innes
4961bf72af updates 2016-10-10 23:04:26 +01:00
Mike J Innes
45d30312b6 tf flatten 2016-10-04 22:50:20 +01:00
Mike J Innes
bf04b70ad1 Float32 by default 2016-10-04 22:36:56 +01:00
Mike J Innes
9e9c57d49b more TF support 2016-10-04 22:23:37 +01:00
Mike J Innes
cc1ca4c3c2 Conv2D tweaks 2016-10-04 22:23:26 +01:00
Mike J Innes
c709041d73 forward useful size method 2016-10-04 22:23:10 +01:00
Mike J Innes
2609d47ce9 work more nicely with TF batching 2016-10-04 21:10:50 +01:00
Mike J Innes
8335ab8134 sort-of working mnist example 2016-09-29 21:28:53 +01:00
Mike J Innes
a2aade718d get basic training working 2016-09-29 20:50:43 +01:00
Mike J Innes
9e35bcd4b7 remove demo code 2016-09-28 17:15:52 +01:00
Mike J Innes
74dc86bfbf backward pass shim 2016-09-28 17:15:41 +01:00
Mike J Innes
b662df6ce1 basic tf backend 2016-09-27 02:16:49 +01:00
Mike J Innes
df38a89d9a some attempts to get mxnet working 2016-09-26 21:44:53 +01:00
Mike J Innes
20c2017d15 fix matmul 2016-09-26 19:42:24 +01:00
Mike J Innes
bec7219a93 awful hack to get both examples working 2016-09-06 18:42:08 +01:00
Mike J Innes
62ede8cd80 use Juno progress bar 2016-09-06 18:37:39 +01:00
Mike J Innes
781f5c0c24 probably-not-working concatenation 2016-09-06 18:11:15 +01:00
Mike J Innes
8961b4c10f basic convnet example working 2016-09-06 18:11:15 +01:00
Mike J Innes
205e1215d6 remove anon model compilation 2016-09-06 18:11:15 +01:00
Mike J Innes
467e829b64 tweaks 2016-09-06 18:11:15 +01:00
Mike J Innes
6503496c39 improve printing 2016-09-06 18:11:14 +01:00
Mike J Innes
abcb6d6351 recurrence proof of concept 2016-09-06 18:11:14 +01:00
Mike J Innes
d58fefb972 tweak note 2016-09-06 18:11:14 +01:00
Mike J Innes
cd968af228 basic unrolling 2016-09-06 18:11:14 +01:00
Mike J Innes
19b5e8bd21 loop lifting 2016-09-06 18:11:14 +01:00
Mike J Innes
edf69ac968 handle recurrence 2016-09-06 18:11:14 +01:00
Mike J Innes
c55f955f1e nicer mxmodel printing 2016-09-06 18:11:14 +01:00
Mike J Innes
2b88be1c7f whoops 2016-09-06 18:11:14 +01:00
Mike J Innes
821efb0cf6 show accuracy less frequently 2016-09-06 18:11:14 +01:00
Mike J Innes
c92cff5dce a bunch of stuff 2016-09-06 18:11:05 +01:00
Mike J Innes
29aab1e4e0 more activation functions 2016-09-06 18:10:21 +01:00
Mike J Innes
526165c897 mxmodel tweaks 2016-09-06 18:10:21 +01:00
Mike J Innes
afac5d8bfe better default init 2016-09-06 18:10:21 +01:00
Mike J Innes
fd67383494 don't print reams of data 2016-09-06 18:10:21 +01:00
Mike J Innes
9675dc765d backprop for sigma 2016-09-06 18:10:21 +01:00
Mike J Innes
e17d1cbe7a fix signs 2016-09-06 18:10:20 +01:00
Mike J Innes
8224c77f7d fix for new api 2016-09-06 18:10:20 +01:00
Mike J Innes
5ea18ddffd revive basic train code 2016-09-06 18:10:20 +01:00
Mike J Innes
6e5e532cc1 update step for mxnet models 2016-09-06 18:10:20 +01:00
Mike J Innes
c1d85abfc2 i fluxed up 2016-09-06 18:10:20 +01:00
Mike J Innes
8c7e74bf9f implement mxnet backward pass 2016-09-06 18:10:20 +01:00
Mike J Innes
2635283bf1 small reorg 2016-09-06 18:10:20 +01:00
Mike J Innes
b8565a4cc3 update api 2016-09-06 18:10:20 +01:00
Mike J Innes
6808a92793 anonymous models 2016-09-06 18:10:20 +01:00
Mike J Innes
51d14cef20 make sure package loads 2016-09-06 18:10:20 +01:00
Mike J Innes
dcdc5fd9c3 rename ModelInput 2016-09-06 18:10:20 +01:00
Mike J Innes
545d4480ed tweaks 2016-09-06 18:10:20 +01:00
Mike J Innes
20b13b9a41 fix 2016-09-06 18:10:20 +01:00