6299f310c7
Experimental feature: creating Storages from raw pointers, in Lua.
2013-03-25 23:08:52 -04:00
e5342e8f86
Experimental feature: creating Storages from raw pointers, in Lua.
2013-03-25 23:08:52 -04:00
914ce58eff
Merge branch 'prettyprint'
2013-02-13 20:54:56 -05:00
64047da4c9
Added a TTY detector.
2013-02-13 13:58:36 -05:00
c1ca8b73a0
add documentation for file:referenced, file:isReferenced, torch.expand, torch.expandAs, torch.repeatTensor
2013-02-13 10:48:26 +00:00
164c2e902f
add repeatTensor function, similar to matlab repmat
2013-02-13 07:58:02 +00:00
5474bf3e74
add isReferenced function to querry the referenced state
2013-02-12 14:16:04 +00:00
45326da58f
reverse logic on referenced
2013-02-12 10:39:41 +00:00
f409840211
add referenced function to disable/enable indexing/referencing behaviour of the writeObject mechanism.
2013-02-12 10:37:23 +00:00
e098ae1382
make tensor.expand create a tensor from correct type, not default type.
2013-02-11 16:50:16 +00:00
6cbdbf4583
DiskFile: removed restrict keyword
2013-02-06 22:52:23 +01:00
a3289c4f51
DiskFile: removed restrict keyword
2013-02-06 22:52:23 +01:00
88b0078383
THDiskFile: workaround insane Mac OS X fread bug ID# 6434977.
2013-02-06 11:27:39 +01:00
f028399417
THDiskFile: workaround insane Mac OS X fread bug ID# 6434977.
2013-02-06 11:27:39 +01:00
47a9334ec1
Merge pull request #100 from jucor/LinspaceFix
...
Fix boundary cases in logspace() and linspace()
2013-01-09 03:54:46 -08:00
b9cb4174a5
Reported the full traceback on a test failure.
2013-01-09 11:32:22 +00:00
a2e455f7a2
Fix boundary cases in logspace() and linspace()
...
For any a < b:
torch.linspace(a, a, 1) returns a 1-dim tensor of a, instead of nan
torch.linspace(a, b, 1) throws an error, instead of returning nan
torch.logspace(a, a, 1) returns a 1-dim tensor of 10^a, instead of nan
torch.logspace(a, b, 1) throws an error, instead of returning nan
torch.test() checks the 4 asserts above
2013-01-08 17:48:26 +00:00
406afa3877
Fix boundary cases in logspace() and linspace()
...
For any a < b:
torch.linspace(a, a, 1) returns a 1-dim tensor of a, instead of nan
torch.linspace(a, b, 1) throws an error, instead of returning nan
torch.logspace(a, a, 1) returns a 1-dim tensor of 10^a, instead of nan
torch.logspace(a, b, 1) throws an error, instead of returning nan
torch.test() checks the 4 asserts above
2013-01-08 17:48:26 +00:00
756083d08a
protect tester calls with pcall.
2013-01-03 10:31:00 +00:00
65f0ea7870
Merge pull request #97 from jucor/RangeEqualBounds
...
Allow for equal bounds in range
2012-12-24 07:49:34 -08:00
e192a14372
Merge pull request #94 from rosejn/force_write
...
Added an optional force option to writeObject.
2012-12-21 04:34:46 -08:00
1fcc8aa31a
make areTablesEqual local.
2012-12-20 17:38:10 +00:00
c6348928d2
Allow for equal bounds in range
2012-12-12 14:35:55 +00:00
4de4a301c5
Allow for equal bounds in range
2012-12-12 14:35:55 +00:00
1aab740694
Allow negative step with range
2012-12-12 13:04:21 +00:00
526def20a0
Allow negative step with range
2012-12-12 13:04:21 +00:00
cc0fd59190
revert the bug fix. fix was buggy too...
2012-12-10 15:29:57 +00:00
f0203f5781
revert the bug fix. fix was buggy too...
2012-12-10 15:29:57 +00:00
8bc0ebdba5
Added an optional force option to writeObject.
...
With this patch you can force writeObject to disregard caching and always write
the given object.
f:writeObject(sample, true)
If only there were javadoc style comments, I would add to the comment string as
well...
2012-12-10 15:07:18 +00:00
68ab6871eb
correction for MACOSX bug mentioned here. https://discussions.apple.com/thread/3413139?start=15&tstart=0
2012-12-09 21:58:51 +00:00
60a232cfbf
correction for MACOSX bug mentioned here. https://discussions.apple.com/thread/3413139?start=15&tstart=0
2012-12-09 21:58:51 +00:00
35e31bd331
Compare tables recursively
2012-12-09 18:30:42 +00:00
956b6643ea
Fix comparison of flat tables
2012-12-09 18:27:13 +00:00
a62907419b
Detect and unit-test a bug in AssertTableEq
2012-12-09 17:44:56 +00:00
ffba0dfbfa
Unit test assertTensorEq and assertTensorNe
2012-12-09 17:44:56 +00:00
d9a46534aa
Document the three new asserts(), and add unit test for TestAssertError()
2012-12-07 20:10:29 +00:00
f361ec16b6
Add three asserts to the Tester class
...
* assertError() expects an error to occur
* assertTensorNe() expects the content of two tensors to differ
* assertTableNe() expects the content of two tables to differ
2012-12-07 20:09:33 +00:00
87a5096293
Fix error message for TensorEQ and TableEQ
2012-12-07 20:09:10 +00:00
8e1e9943a4
Document the extended parameter range for Bernoulli
2012-12-03 14:25:26 +00:00
6bbcc99187
Allow p=0 and p=1 for Bernoulli distribution
2012-12-03 14:23:12 +00:00
8939a5744a
Allow p=0 and p=1 for Bernoulli distribution
2012-12-03 14:23:12 +00:00
f49e22cff2
Merge remote-tracking branch 'upstream/master'
2012-11-29 10:34:41 +00:00
fe763ec715
count and report assertions
2012-11-27 17:41:32 +00:00
7bd8b98124
Fixed doc typo.
2012-11-19 19:03:15 +00:00
8ed97e6961
Using a local var.
2012-10-31 12:26:30 +00:00
ba02f818b5
Upgraded CUDA code to compute capability 2.0 (48kB of shared mem)
2012-10-27 15:53:50 -04:00
526df4b3c0
Merge pull request #71 from akfidjeland/cuda-maths
...
Added missing form of CudaTensor:add
2012-10-25 10:03:25 -07:00
f78dbc1da5
fix ATLAS detection
2012-10-22 22:28:19 +02:00
6bd1bbb04b
fix ATLAS detection
2012-10-22 22:28:19 +02:00
781d35d251
max/min/sort: no need to increment optional IndexTensor
2012-10-22 21:55:57 +02:00
