- 17 12月, 2014 4 次提交
-
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
The NoContext object in graph will serve a "absent context" marker.
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
This depends on the Op that is in the node and context is optional.
-
- 16 12月, 2014 10 次提交
-
-
由 Frédéric Bastien 提交于
Add print during unit test to help preventing timeouts
-
由 Frédéric Bastien 提交于
Multi fixes (pieces of the multi-gpu code)
-
由 Arnaud Bergeron 提交于
-
由 Pierre Luc Carrier 提交于
-
由 Frédéric Bastien 提交于
Fix buildbot error from #2329
-
由 Arnaud Bergeron 提交于
I'm not sure what this test is expecting. It seems to test that the gpu will magically work and optimize the graph even though we disable InputsToGpuOptimizer (which is obviously not the case). I wonder why it passed before. I wonder if this is the right fix. I also wonder what is it testing exactly. In any case it seems to pass now.
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
It is suitably unique and, more importantly, shared with the rest of the methods.
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
-
- 15 12月, 2014 2 次提交
-
-
由 Frédéric Bastien 提交于
Scan push out sum of dot
-
由 Pierre Luc Carrier 提交于
-
- 14 12月, 2014 4 次提交
-
-
由 Frédéric Bastien 提交于
Convert SetSubtensor to IncSubtensor when possible
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
-
- 13 12月, 2014 1 次提交
-
-
由 Arnaud Bergeron 提交于
This did come up in an implementation of adagrad at some point and will be much faster on GPU.
-
- 12 12月, 2014 1 次提交
-
-
由 abergeron 提交于
Do not move the GpuSubtensor to the GPU when this do not make sence
-
- 11 12月, 2014 2 次提交
-
-
由 Frédéric Bastien 提交于
[BUG] Fix bug in {inc,set}_subtensor in advanced1 case -
由 abergeron 提交于
Join opt. Remove empty constant as inputs and merge MakeVector inputs.
-
- 10 12月, 2014 3 次提交
-
-
由 Frederic Bastien 提交于
-
由 Pierre Luc Carrier 提交于
-
由 Frederic Bastien 提交于
-
- 09 12月, 2014 9 次提交
-
-
由 Frederic Bastien 提交于
do the join change for gh-2328
-
由 Pierre Luc Carrier 提交于
-
由 Frederic 提交于
-
由 Frederic 提交于
fix gh-2327
-
由 Pascal Lamblin 提交于
Remove useless warning and make the graph more optimized in old supported cases.
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic 提交于
When we do this on an inputs/shared var and that this is the only use of it in the graph, we should only transfer to the GPU the small computation.
-
- 08 12月, 2014 1 次提交
-
-
由 Pascal Lamblin 提交于
-
- 07 12月, 2014 2 次提交
-
-
由 Pascal Lamblin 提交于
-
由 Pascal Lamblin 提交于
Reported by Stanislas Lauly
-
- 06 12月, 2014 1 次提交
-
-
由 Pierre Luc Carrier 提交于
-