- 03 11月, 2016 3 次提交
-
-
由 A. Unique TensorFlower 提交于
Change: 137959663
-
由 Dan Smilkov 提交于
Change: 137954868
-
由 A. Unique TensorFlower 提交于
Change: 137952318
-
- 02 11月, 2016 37 次提交
-
-
由 Peter Hawkins 提交于
Perform constant folding in the calling thread, rather than using a singleton threadpool. Avoids a potential deadlock when constant folding is called in a re-entrant manner. Change: 137945911
-
由 Nikhil Thorat 提交于
-Update the help text to not use "bookmarks" as a noun and a verb -Change the download icon to the save icon -Disable the save icon if there are no bookmarks -Change visible nomenclature to save and load instead of upload/download -Fix the radio button edge-click bug that selects the first bookmark -Do some basic checks on the shape of the dataset and the bookmarks. Change: 137943270
-
由 Tim Harley 提交于
This fixes a race condition in coordinator.join(). Removed workaround in QueueRunnerTest. Change: 137933187
-
由 A. Unique TensorFlower 提交于
-
由 Gunhan Gulsoy 提交于
Change: 137917462
-
由 Gunhan Gulsoy 提交于
Change: 137914787
-
由 Shanqing Cai 提交于
Change: 137908539
-
由 Justine Tunney 提交于
Change: 137907844
-
由 Dan Smilkov 提交于
Change: 137905523
-
由 Dan Mané 提交于
Specifically, suppose you have a line chart with 30 runs. If you mouse over, you will have 30 tooltips, which is hard to read. Previously, if you zoom in so that only 1 run is visible on the screen, you would still get 30 tooltips. Now, if you mouse over you will only see tooltips for runs that are actually on the screen. Change: 137905115
-
由 Alexey Surkov 提交于
This is now consistent with what other file systems have. Change: 137892031
-
由 A. Unique TensorFlower 提交于
Change: 137889948
-
由 Gunhan Gulsoy 提交于
Change: 137888607
-
由 A. Unique TensorFlower 提交于
specify additional tf_kernel_library targets that should be grouped into the omnibus cc_library target that includes all generated rules, but should not be considered dependencies of each individual generated tf_kernel_library target. Use the new attribute to split out some existing deps entries and avoid unnecessary transitive dependencies. Split out the :l2loss_op library so that it doesn't share the same deps list with the rest of the "nn" libraries, in particular avoiding a dependency on :pooling_ops which in turn depends on the expensive :eigen_helpers. Likewise, split out :lookup_table_{init_}op from the "data_flow" libraries. Change: 137887998
-
由 Eugene Brevdo 提交于
Change: 137887950
-
由 A. Unique TensorFlower 提交于
Added a button below each image in the Tensorboard image panel. Clicking this new button shows an actual size version of the image. Change: 137886562
-
由 A. Unique TensorFlower 提交于
Change: 137885496
-
由 Jonathan Hseu 提交于
in TensorBoard. - Fixes GCS so that it still includes prefixes even if items is empty. - Change list_directory to use GetChildren() instead of GetMatchingFiles. GetMatchingFiles in GCS has a quirk where, unlike other filesystems, it doesn't include directories. There's no obvious fix given the GCS API. Change: 137885363
-
由 Gunhan Gulsoy 提交于
Change: 137883577
-
由 Gunhan Gulsoy 提交于
Change: 137881303
-
由 Alexey Surkov 提交于
In GCS, a folder 'gs://bucket/path' can contain a zero-length object 'gs://bucket/path/', which primarily exists to make IsDirectory('gs://bucket/path') return 'true'. Naturally GetChildren returned the directory marker for the specified directory whenever it existed, represented as an empty string, which violates assumptions that people normally have from this functions. Change: 137878839
-
由 Zongheng Yang 提交于
It's a weak test + causes complications if we consider both V1 & V2. Change: 137878590
-
由 Dan Smilkov 提交于
Change: 137877854
-
由 A. Unique TensorFlower 提交于
nodes and use the information provided by them to set the min/max values on quantization-related nodes. In eightbit mode, also changed how constant weights are quantized - instead of doing it as a step after the main recursion, do it during the main recursion. This allows the float inputs to FakeQuantWithMinMaxVars to be excluded from quantization. In eightbit mode, maintain more state in the stack during recursion. Also change quantize reshape registration to register always and not use TF_CALL_xyz; this matches other quantized ops. Change: 137877226
-
由 A. Unique TensorFlower 提交于
Change: 137866950
-
由 Yuefeng Zhou 提交于
Change: 137866409
-
由 Joshua V. Dillon 提交于
Change: 137863303
-
由 A. Unique TensorFlower 提交于
Change: 137863207
-
由 Alexey Surkov 提交于
Change: 137861834
-
由 A. Unique TensorFlower 提交于
Change: 137855838
-
由 Illia Polosukhin 提交于
Change: 137854718
-
由 A. Unique TensorFlower 提交于
For all but the edge tiles, we can unroll the loops that read and write from our shared memory tile. Change: 137853968
-
由 A. Unique TensorFlower 提交于
SparseTensor.from_value. Change: 137850511
-
由 Illia Polosukhin 提交于
Change: 137850119
-
由 A. Unique TensorFlower 提交于
Change: 137849675
-
由 Patrick Nguyen 提交于
Change: 137849082
-
由 A. Unique TensorFlower 提交于
Change: 137848991
-