| Commit message (Collapse) | Author | Age |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fix issue of tf.gfile.Exists with Unicode on Windows
This fix tries to address the issue raised in 17695 where
tf.gfile.Exists with Unicode on Windows does not work as expected.
The issuse comes from the usage of `_access` (vs. _waccess).
This fix adds a test and fixes the issue.
This fix fixes 17695.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Replace _access with _waccess on Windows.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
|
|
|
|
|
|
| |
* Fix the list format in using_tpu.md
* Fix the indent before tf.contrib.cluster_resolver.TPUClusterResolver
|
| |
|
|
|
|
|
|
|
|
| |
* Fix link
* fix link to use the "document" syntax.
the api_guide pages are counted as "documents" not api pages, so they use the document syntax [described here](https://www.tensorflow.org/community/documentation#links_in_markdown).
|
|\
| |
| | |
Branch 188893722
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Fix warning in embedding_ops
This fix fixes the warning in embedding_ops.py by switching
from keep_dims to keepdims for tf.reduce_prod
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Fix warning in contrib/learn/python/learn/ops/embeddings_ops.py
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
In recurrent_quickdraw.md, the link `${tf.train.Example}` is
not rendered correctly. This fix fixes the broken link with
correct rendering. Now the link is rendered the same way
as tf.train.Example in api_guides/python/reading_data.md and extend/new_data_formats.md
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Fix format issue in ffmpeg
This fix tries to address the issue raised in 17533 where
some ouput generated by ffmpeg may not be parsed correctly.
There are two issues in the ffmpeg parser:
1. The size of the frame could be `rgb24, 640x360 [SAR 1:1 DAR 16:9]`,
with extra `[...]` after `640x360`.
2. The number of frames could be `frame=12488` or `frame= 166`, with or
without the sapce (` ` in between `frame=` and the number).
This fix fixes the parser issues in ffmpeg.
This fix fixes 17533.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Fix ffmpeg parser for `frame= 166 ` or `frame=12488 `
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
|
| |
| |
| | |
espected -> expected
|
| |
| |
| | |
Alphabetized races: standard procedure in documentation.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Improvements for Android TV
To build for Android TV, user need to modify the orientation from "portrait" to "landscape"
for the four activities in tensorflow/examples/android/AndroidManifest.xml
1. Show a toast when no camera is detected
2. Game Controller L1 button and remote controller DRAP_Center also triggers debug output
3. Fix the layout of Stylize app on Android TV. The style selector will be on right side
4. User can use DPAD to navigate through styles in Stylize app
* Fix README grammar, keyDown pass down, and unused import
* Calculate style selector width and num of style in a row programmatically
* Fix imports
|
| |
| |
| |
| |
| |
| |
| |
| | |
(#17594)
* Fix the broken link of tf-learn's iris tutorial and also fix some messed up format and typo
* revert a wrong typo fix
|
| |
| |
| |
| |
| |
| | |
* Fix broken link in programmers_guide/faq and some minor format
* revert the change of The * is an asterisk denoting a footnote, which is referenced at the bottom on line 296
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Fix the broken links in tutorial/layers
* revert link which is already working with - split title
* correct new lines leads to broken link
* revert changes to keep existing anchors
|
| | |
|
|/| |
|
| |
| |
| |
| | |
PiperOrigin-RevId: 188893722
|
| |
| |
| |
| | |
PiperOrigin-RevId: 188890943
|
| |
| |
| |
| |
| |
| | |
FisherEstimator's constructor. This allows layers and losses to be registered after the FisherEstimator (or KFACOptimizer) has been constructed.
PiperOrigin-RevId: 188889252
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This will allow you to run it with:
`python -m tensorflow.python.tools.import_pb_to_tensorboard ...`
Without any additional downloads or setup.
see #8854
PiperOrigin-RevId: 188888501
|
| |
| |
| |
| | |
PiperOrigin-RevId: 188880419
|
| |
| |
| |
| | |
PiperOrigin-RevId: 188880237
|
|\ \
| | |
| | | |
Fix the script entry point for freeze_graph.
|
| | |
| | |
| | |
| | |
| | |
| | | |
We allow now to fuse (A dot B) times a constant, which is available in cublas.
PiperOrigin-RevId: 188878210
|
| | | |
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 188828898
|
| | | |
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 188828898
|
| | | |
|
|\ \ \
| | | |
| | | | |
Branch 188738133
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 188828898
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
path (#17605)
* MKL DNN: fix the TF1.6 speed issue by fixing MKL DNN LRN
* fixed typos in the doc for LrnRewrite
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
exposes sequence methods under tf.contrib.feature_column.
PiperOrigin-RevId: 188826187
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 188824362
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 188824058
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 188817976
|
|\ \ \ \
| | | | |
| | | | | |
contrib/quantize: minor spelling
|
| | | | |
| | | | |
| | | | |
| | | | | |
PiperOrigin-RevId: 188817194
|
|\ \ \ \ \
| | | | | |
| | | | | | |
Add scan command to saved_model_cli to check for security sensitive ops.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
PiperOrigin-RevId: 188815493
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
PiperOrigin-RevId: 188812445
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
PiperOrigin-RevId: 188807786
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
PiperOrigin-RevId: 188803724
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
ExtractConstantSubgraph doesn't copy control edges, which are sometimes necessary
to correctly evaluate conds (at the very least). Avoid evaluating conds at all to
address this.
PiperOrigin-RevId: 188803649
|
| |/ / / /
|/| | | | |
|