提交 · 0a531470b3d5dbbcff3cbfa26250630cc88ca22e · testgroup / pytensor

25 10月, 2016 3 次提交
- Don't test sum with a mix of complex and scalar. · 0a531470
  由 Arnaud Bergeron 提交于 10月 24, 2016
```
Sum with a complex input and scalar output doesn't work because of the
ambiguity of casting complex to scalar.  The reverse doesn't work
because of the ambiguity of the casting in the gradient.  Therefore
disallow both.
```
  0a531470
- Fix typo. · 7da49941
  由 Arnaud Bergeron 提交于 10月 24, 2016
  
  7da49941
- Add mission semicolon. · 2af0f5e5
  由 Arnaud Bergeron 提交于 10月 24, 2016
  
  2af0f5e5
21 10月, 2016 3 次提交
- Apparently this was missing. · 7f154d35
  由 Arnaud Bergeron 提交于 10月 20, 2016
  
  7f154d35
- Make sure the reference is not float16. · 2cb0c380
  由 Arnaud Bergeron 提交于 10月 20, 2016
  
  2cb0c380
- Change the test to expect regular numpy behaviour. · c14c5957
  由 Arnaud Bergeron 提交于 10月 20, 2016
  
  c14c5957
20 10月, 2016 7 次提交
- Don't allow operations that are deprecated in numpy. · fe500996
  由 Arnaud Bergeron 提交于 10月 19, 2016
  
  fe500996
- Don't check the .type of a string. · c6b619c9
  由 Arnaud Bergeron 提交于 10月 19, 2016
  
  c6b619c9
- flake8 · 6058586d
  由 Arnaud Bergeron 提交于 10月 19, 2016
  
  6058586d
- Make sure the scalar ops handle bool correctly and don't produce irregular booleans. · b3a291fa
  由 Arnaud Bergeron 提交于 10月 19, 2016
  
  b3a291fa
- Merge pull request #21 from lamblin/bool · 33bd55ef
  由 abergeron 提交于 10月 19, 2016
```
Fix tests for CAReduceDtype with bool
```
  33bd55ef
- flake8 · c25f2cc0
  由 Pascal Lamblin 提交于 10月 19, 2016
  
  c25f2cc0
- Make int64 default dtype for bool reduction · ed44bb59
  由 Pascal Lamblin 提交于 10月 19, 2016
```
This is consistent with numpy
```
  ed44bb59
19 10月, 2016 10 次提交
- Don't change the printout. · 91333f00
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  91333f00
- Scalar needs __init__. · 450ce1b2
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  450ce1b2
- Make sure we can load and work with old pickles. · f896baea
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  f896baea
- flake8 for doc (it's cleaner). · ba2305e6
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  ba2305e6
- Flake8 · d804254d
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  d804254d
- No fix needed anymore. · 9b6a964a
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  9b6a964a
- Don't use neg for unsigned types. · 4aa42cc3
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  4aa42cc3
- Only allow bitwise ops between ints. · 5ef9904f
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  5ef9904f
- Fix problems in sanbox/cuda. · 376ca250
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  376ca250
- Fix tests in gpuarray for bool. · 405ccc5d
  由 Arnaud Bergeron 提交于 10月 18, 2016
  
  405ccc5d
18 10月, 2016 8 次提交
- Flake8 · 8ca5b3eb
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  8ca5b3eb
- Use integer_dtypes if we only want integers. · f0d21787
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  f0d21787
- Add a integer_{types,dtypes} which is only integer, no bools. · 0fe88fc6
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  0fe88fc6
- Remove BinCountOp since it's deprecated and the test breaks with bool. · b1b835dd
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  b1b835dd
- Minimum is numpy 1.9.1 · dd288e88
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  dd288e88
- No docstrings for tests. · 0f4be8d9
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  0f4be8d9
- Bump cache version for comparison ops since their return type changed. · db809a2a
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  db809a2a
- Flake8 + review fixes. · 99012e14
  由 Arnaud Bergeron 提交于 10月 17, 2016
  
  99012e14
14 10月, 2016 7 次提交
- Change mean() to always return the same dtype even if axis=[]. · a476757a
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  a476757a
- Remove line continuations. · 872f34a3
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  872f34a3
- Fix almost all the tests. · f0a58c50
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  f0a58c50
- Fix reduction test to actually iterate through all the dtypes. · 12b74a54
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  12b74a54
- Remove redundant definitions of *_dtypes. · 9bddfc4f
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  9bddfc4f
- Make Any and All return bools. · ac8ce341
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  ac8ce341
- Make bool part of the discrete types. · 7563a320
  由 Arnaud Bergeron 提交于 10月 13, 2016
  
  7563a320
13 10月, 2016 2 次提交
- Switch subtensor check to bool and fix some things in tensor. · 1661f0bc
  由 Arnaud Bergeron 提交于 10月 12, 2016
  
  1661f0bc
- Add stuff so that scalar tests pass with the new bool type. · 27654022
  由 Arnaud Bergeron 提交于 10月 12, 2016
  
  27654022