add2896ae4
Not hard-coding build locations
2018-07-09 08:37:44 -07:00
e8f2db45e4
Add bool in/out support to pass ONNX backend tests.
2018-07-06 21:50:53 -07:00
624bf7d82b
String changes to support conversion from ASCII, UCS2, UCS4, UTF8, UTF16, and UTF32 strings
2018-07-06 18:48:27 -07:00
ed9de4713e
Merge branch 'yanchen/mpi-makefile'
2018-07-07 00:54:06 +00:00
c11f51dbfd
Merge branch 'sedymche/onnx-backend'
2018-07-07 00:29:50 +00:00
2626ae7ff4
Merge branch 'master' into yanchen/mpi-makefile
2018-07-06 14:36:58 -07:00
70cf3a39bc
Merge branch 'master' into yanchen/docker-copy
2018-07-06 14:36:37 -07:00
7b3a840ca4
Update skip list for ONNX backend tests.
2018-07-06 13:20:39 -07:00
afbe91e39a
update NoGPU.cpp
2018-07-06 11:08:36 -07:00
19ffa068bd
Merge branch 'master' into bowbao/gather_grad
2018-07-06 10:03:13 -07:00
6adce79408
Merge branch 'master' into bowbao/gather_grad
2018-07-06 10:01:44 -07:00
10c1d6b303
nit
2018-07-06 09:54:26 -07:00
3bfa12b9da
Merge branch 'master' of https://github.com/Microsoft/CNTK into tix/fix_netcoreTestFailure
2018-07-05 21:49:20 -07:00
0b317eb88b
Merge branch 'bowbao/clone_segfault'
2018-07-06 02:04:49 +00:00
92796cfd8b
fixed bug around mapping from columnValidMask to data. Instead of a 1 to 1 mapping, each element in columnValidMask maps to a grid of multiple rows and columns in data matrix.
2018-07-05 18:37:40 -07:00
04844b0664
Fixed path
2018-07-05 18:22:25 -07:00
610abaf3da
Added comment
2018-07-05 16:59:38 -07:00
a89a334cf5
Fixed path
2018-07-05 16:55:33 -07:00
dc29eb3057
Enabled the test for py35 so that it gets run by default in bvt
2018-07-05 16:52:10 -07:00
6cc772f693
small change for more reuse of code
2018-07-05 16:42:51 -07:00
56d0ada718
remove the data folder created by the python script
2018-07-05 15:46:41 -07:00
6e265dae47
Fixed an issue for copying patches
...
Unlike Linux-like cp instruction, docker COPY instruction doesn't
preserve the top-level dir structure (e.g. it copies the contents
of the top-level dir rather than the dir itelse to the image).
Our Dockerfile for generating GPU image has correct behavior though.
2018-07-05 14:42:02 -07:00
9a075b356c
Merge branch 'master' into bowbao/clone_segfault
2018-07-05 14:40:44 -07:00
0df5c39fbb
Add method ScatterToIndicesWithMask for class Matrix(both CPUMatrix and GPUMatrix).
...
Change GatherNode Backprop to use the new method instead of altering data for gaps in input matrix.
2018-07-05 14:20:13 -07:00
49e35831bc
Copy onnx_cntk from https://github.com/onnx/onnx-cntk .
2018-07-05 14:13:40 -07:00
9667426382
Fixing run-test to grab for the correct dll for netcore build
2018-07-05 12:12:36 -07:00
ed9f8796aa
Merge branch 'master' into yanchen/mpi-makefile
2018-07-05 10:00:01 -07:00
c1430bff1c
added fmeasure loss function
2018-07-05 20:58:27 +08:00
17e6a012f9
Merge branch 'pull/3277'
2018-07-04 03:49:50 +00:00
7c74244387
set missing value back to 0 in case of unexpected usage.
2018-07-03 18:48:01 -07:00
1e058cedcf
fix Gather op's incorrect gradient value.
...
* the error was due to that we pad 0 as default value for missing gaps. All these then each contribute 1 to the gradient of reference at index 0. The fix is to mask missing values in indices matrix to negative, and in Matrix scatter implementation to check and skip negative indices. (previous Matrix CPU implementation already checks for negative indices)
2018-07-03 18:34:11 -07:00
c049493681
Merge branch 'master' into yanchen/mpi-makefile
2018-07-03 16:07:22 -07:00
81ced59adb
add None check for python clone substitution.
...
* it seems None somehow escapes the typecheck when being converted to class CNTK::Variable thus causing the crash.
2018-07-03 15:19:33 -07:00
8577222563
Merge branch 'tix/convertEvalExampleTestToNetCore' of https://github.com/Microsoft/CNTK
2018-07-03 15:02:22 -07:00
25a0896e99
use system CXX for building Multiverso on Linux
...
We used to set CXX to mpic++, which was then used for building
Multiverso. Unfortunately, this kind of configuration conflicted
with find_package(MPI), and hence caused failure.
This commit fixed the issue by using system CXX for building
Multiverso. It also fixed two other issues:
* EVAL_LIB depends on libmultiverso, so we had to set
MULTIVERSO_LIB before the EVAL_LIB rule
* cmake variable doesn't have a "PATHNAME" type, we used
PATH instead. We also replace FILEPATH with PATH because
BOOST_LIBRARY_DIRS points to a path rather than a file.
2018-07-03 13:55:57 -07:00
dd9705bfec
Fixing LRN op to match ONNX 1.2 spec. Size attribute has semantics of diameter and not radius, as was the case before.
2018-07-03 12:40:12 -07:00
8d393227a8
Converted eval example to use netcore
2018-07-02 18:33:08 -07:00
63603e0f70
SqueezeNet numbers mistmatch fix (due to an issue on softmax layer)
2018-07-02 13:58:49 -07:00
07fd96f357
support ONNX1.2.2 cast op
2018-07-02 08:38:50 -07:00
2f91f7463b
Fix saved model name output
...
Corrected the output showing the name of the model being saved
2018-07-01 23:53:29 +10:00
bece037f63
Merge branch 'sedymche/onnx-trig'
2018-06-29 01:52:40 +00:00
69938f2992
Adding full support for 'alpha' attribute in ELU op in ONNX and CNTK.
2018-06-28 15:34:30 -07:00
e95f92cd5e
Add Tan/Atan ops to CNTK (with ONNX support).
2018-06-28 14:03:02 -07:00
c589619dec
update to ONNX1.2.2
2018-06-28 08:38:36 -07:00
e160304efc
Merge branch 'sptiwari/onnx_avg_pool_stride'
2018-06-28 04:02:28 +00:00
6340792c67
Merge branch 'zhujie/straightthrough_unaryop'
2018-06-28 00:55:55 +00:00
2e2962547c
Merge branch 'tix_fixRestoredTestFailures' of https://github.com/Microsoft/CNTK
2018-06-27 15:25:50 -07:00
805c4a6ffe
Fixing strides and dilation attributes in ONNX AveragePool, MaxPool, Conv, ConvTranpose import.
2018-06-27 14:54:47 -07:00
3c389ed28d
fixing a bug in straight through implementation
2018-06-27 11:44:49 -07:00
cc368aea82
fixing a bug in backwarod propagation of straightthrough
2018-06-27 11:44:49 -07:00
David Brownell
Sergii Dymchenko
David Brownell
Yang Chen
Bowen Bao
TJ
delzac
Spandan Tiwari
Peyman Manikashani
liqfu
jediknight31
Jie Zhu