- 21 6月, 2016 5 次提交
-
-
由 A. Unique TensorFlower 提交于
Change: 125375538
-
由 A. Unique TensorFlower 提交于
comment/reformat in examples. Change: 125374843
-
由 A. Unique TensorFlower 提交于
Refactor Get2dOutputSizes/Get2dOutputSizesVerbose/Get3dOutputSizes to share a common 1-dimensional GetWindowedOutputSize/GetWindowedOutputSizeVerbose. The output sizes and padding of each dimension of a windowed operation (such as convolution or pooling) are orthogonal and can be computed independently. We can simplify the code by providing a 1D size computation and calling it for each dimension. Also remove special cases for 1x1 spatial convolutions in dimension calculations; they add complexity and are a case that the general code handles correctly. In general, 2D convolutions and their gradients have a lot of shape calculation code that is duplicated for each spatial dimension. This CL is a step in the direction of treating spatial dimensions the same so we can share more code. Change: 125360639
-
由 Sherry Moore 提交于
Change: 125359897
-
由 Derek Murray 提交于
This change should fix issues where (e.g.) a node is added after queue runners are started, and an error like `NotFoundError: FetchOutputs node input_producer_Dequeue:0: not found` is reported. The cause of the error was an incorrect multiple read of the `Graph.version` property. A graph has a `version` property that is (currently) defined to be the count of nodes in the graph. To keep the front-end and session graphs in sync, the `tf.Session` maintains a `_current_version` property, which is intended to be the last version at which the graph was synced to the backend (via TF_ExtendGraph). However, the logic was wrong in the face of concurrent writes to the graph: 1. x = read `Session._current_version`. 2. y = read `Session._graph.version`. 3. If y > x: - Call `TF_ExtendGraph(Session._graph.as_graph_def(from_version=x))` 4. z = read `Session._graph.version`. 5. Set `Session._current_version` to z. Written this way, the race condition is clear: a node added after step 3 and before step 4 would never be included in an TF_ExtendGraph call, and so would never be added to the graph. The fix involves returning the current graph version atomically with the GraphDef, using a new internal method called `Graph._as_graph_def()`. These are protected by the graph's mutex, to ensure that no node can be added between creating the GraphDef and reading the version. Fixes #2957. Change: 125356853
-
- 18 6月, 2016 16 次提交
-
-
由 David Soergel 提交于
Change: 125246026
-
由 A. Unique TensorFlower 提交于
Change: 125225689
-
由 A. Unique TensorFlower 提交于
Change: 125225345
-
由 Ian Langmore 提交于
Update BaseDistribution API for NaN handling. Doc only change in BaseDistribution (which does not specify __init__ kwargs). Change: 125224922
-
由 A. Unique TensorFlower 提交于
Change: 125215862
-
由 A. Unique TensorFlower 提交于
Change: 125215192
-
由 A. Unique TensorFlower 提交于
Change: 125215108
-
由 A. Unique TensorFlower 提交于
Change: 125208073
-
由 Yuan Yu 提交于
Change: 125206796
-
由 Manjunath Kudlur 提交于
Change: 125205961
-
由 A. Unique TensorFlower 提交于
Change: 125205938
-
由 A. Unique TensorFlower 提交于
Change: 125190689
-
由 Dan Mané 提交于
Change: 125190301
-
由 A. Unique TensorFlower 提交于
Change: 125189009
-
由 A. Unique TensorFlower 提交于
Change: 125180641
-
由 Shanqing Cai 提交于
Change: 125178681
-
- 17 6月, 2016 19 次提交
-
-
由 A. Unique TensorFlower 提交于
Change: 125134620
-
由 A. Unique TensorFlower 提交于
Change: 125133388
-
由 A. Unique TensorFlower 提交于
Change: 125133220
-
由 Derek Murray 提交于
Change: 125132823
-
由 Manjunath Kudlur 提交于
unique_ptr<OpRegistrationData>, so that the OpDef building can be deferred until an call to Lookup. Change: 125132276
-
由 A. Unique TensorFlower 提交于
Change: 125131602
-
由 Zongheng Yang 提交于
Change: 125130833
-
由 Zongheng Yang 提交于
tf.rank() is recently patched to work on SparseTensor. Change: 125127586
-
由 Vijay Vasudevan 提交于
Switches to return a std::unique_ptr<> to convey transfer of ownership. Keeps existing functions temporarily for backwards compatibility. Change file_system.h interface to switch entirely to the new interface, change all callers. If this breaks someone in the public, the interface change should be straightforward. Change: 125127028
-
由 A. Unique TensorFlower 提交于
Change: 125126928
-
由 Yutaka Leon 提交于
Change: 125125512
-
由 A. Unique TensorFlower 提交于
Change: 125120055
-
由 Geoffrey Irving 提交于
Change: 125115298
-
由 Jianmin Chen 提交于
Change: 125114422
-
由 A. Unique TensorFlower 提交于
Change: 125113500
-
由 A. Unique TensorFlower 提交于
Change: 125113173
-
由 A. Unique TensorFlower 提交于
Change: 125112939
-
由 A. Unique TensorFlower 提交于
for comparison, instead of copying to strings. Change: 125112408
-
由 A. Unique TensorFlower 提交于
Change: 125110928
-