aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Standalone benchmark for Tensorflow models that runs on desktop and Android.Gravatar Andrew Harp2016-04-12
| | | | Change: 119706066
* Add a image.Gravatar Jianmin Chen2016-04-12
| | | | Change: 119703236
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119698585
* Adds option to pass callable initializer function to Variable constructor to ↵Gravatar Wei Ho2016-04-12
| | | | | | | allow colocation of variable initialization with the device the variable is on, instead of always being on the chief supervisor. Also updates variable_scope.get_variable() and create_partitioned_variables() to take advantage of this when an initializer fn is passed in. Change: 119697860
* Avoid using functions from the std namespace on GPU.Gravatar Benoit Steiner2016-04-12
| | | | Change: 119695799
* Remove lib/core/command_line_flags.h, which is superceded by ↵Gravatar Vijay Vasudevan2016-04-12
| | | | | | util/command_line_flags.h Change: 119694323
* Added a unit test to batch normalization.Gravatar Pete Warden2016-04-12
| | | | Change: 119691101
* Extends functionality of Tensor::flat_inner_dims() and ↵Gravatar A. Unique TensorFlower2016-04-12
| | | | | | Tensor::flat_outer_dims() by adding a template param to specify the desired rank of the output. This is useful, e.g., in order to turn a tensor into a (rank-3) batch of matrices. Change: 119685357
* Wrapper for SlimRNNCellsGravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119671640
* More tests and minor fixes for tf/contrib/tensor_forest.Gravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119666432
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119657458
* Modified supervisor event writing behavior to be safer.Gravatar A. Unique TensorFlower2016-04-12
| | | | | | * Set supervisor fields for event writing to None for non-chiefs, so that non-chiefs do not accidentally write summaries. * Added RuntimeErrors for public API endpoints to alert users when they try to use non-chief supervisors to write events. Change: 119654303
* Add unused placeholder for readahead file path support.Gravatar Dan Mané2016-04-12
| | | | Change: 119650236
* test_installation.sh: head command compatibility on MacGravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119649248
* Removing lazy loading of sdca_ops.so since it is no longer necessary.Gravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119648305
* Removing lazy loading of sdca_ops.so since it is no longer necessary.Gravatar A. Unique TensorFlower2016-04-12
| | | | Change: 119643238
* Applying editorial changes to the distributed how-to.Gravatar Derek Murray2016-04-11
| | | | Change: 119605636
* Add SyncReplicasOptimizer test in dist_testGravatar A. Unique TensorFlower2016-04-11
| | | | | | | | | | | | | | Usage example: ./remote_test.sh --num-workers 3 --sync-replicas Also changed: 1) In local and remote tests, let different workers contact separate GRPC sessions. 2) In local and remote tests, adding the capacity to specify the number of workers. Before it was hard-coded at 2. Usage example: ./remote_test.sh --num-workers 2 --sync-replicas 3) Using device setter in mnist_replica.py Change: 119599547
* Enable test_benchmark.* files for Android build.Gravatar Andrew Harp2016-04-11
| | | | Change: 119591021
* Updated the distributed runtime how-to to use the new API in `tf.train.*`.Gravatar Derek Murray2016-04-11
| | | | Change: 119589456
* Redo the TensorBoard tutorial code (mnist_with_summaries.py).Gravatar Dan Mané2016-04-11
| | | | | | | | | Goals: - Have enough of each summary type that tag grouping is useful. (Wound up recording e.g. mean and stddev and min/max for each variable) - Use every summary type (adds images) - Write to multiple directories so there are several "runs" Change: 119585022
* Update released TensorBoard.Gravatar Dan Mané2016-04-11
| | | | | | Update bower dependencies. Also force urls to lowercase. Change: 119584968
* Add stat_summarizer.h into tensorflow/core/BUILDGravatar A. Unique TensorFlower2016-04-11
| | | | Change: 119572994
* Adding link to readme in TensorBoard Toolbar.Gravatar A. Unique TensorFlower2016-04-11
| | | | | Moving favicon to datauri. Change: 119569013
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-04-11
| | | | Change: 119565375
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-04-11
| | | | Change: 119565115
* Restrict the use of OUT_OF_RANGE to what it was intended to.Gravatar A. Unique TensorFlower2016-04-11
| | | | | | | | | | | | Clarify that OUT_OF_RANGE is raised only when reaching the end of input for interable contents. Change the few places where we incorrectly raised OUT_OF_RANGE to raise ILLEGAL_ARGUMENT instead. This will make code that catches the OUT_OF_RANGE exception more robust as it won't get confused by spurious uses of the exception class. Change: 119560848
* Make the TF_CALL_*() macros drop the semicolon, instead putting that burdenGravatar A. Unique TensorFlower2016-04-11
| | | | | | | | | on the caller. This allows us to use the macros for other purposes than calling REGISTER_KERNEL, in particular in variadic template parameter lists. Update REGISTER_KERNEL_BUILDER accordingly to add a semicolon, so that existing code continues to compile. Change: 119551677
* Clean up benchmark.py after previous modifications and add unit test.Gravatar Eugene Brevdo2016-04-11
| | | | Change: 119549296
* Add additional space after "Examples" so ```python``` can renderGravatar A. Unique TensorFlower2016-04-11
| | | | | correctly. Change: 119549145
* Clarification of object ownership in ::tensorflow::EnvGravatar A. Unique TensorFlower2016-04-11
| | | | Change: 119544956
* Add an option to docs.py collect_members to skip certain symbolsGravatar Geoffrey Irving2016-04-11
| | | | | This is used only within Google. Change: 119543426
* Add Dimension.__str__ and improve TensorShape.merge_with errorsGravatar Geoffrey Irving2016-04-11
| | | | | | | | | | | | | | | Dimension.__str__ now spits out ? for unknown dimensions and an integer otherwise. Previously this logic was contained in TensorShape.__str__. In addition, the exception produced TensorShape.merge_with now encodes all of the two shapes, so something like Dimensions 2 and 9 are not compatible becomes Shapes (?, 2) and (4, 9) are not compatible Change: 119536897
* Clarifies some documentation in server_lib.pyGravatar A. Unique TensorFlower2016-04-11
| | | | Change: 119533248
* tensorflow: support usage of eigen thread poolGravatar A. Unique TensorFlower2016-04-11
| | | | | Use eigen ThreadPool instead of tensorflow one if TENSORFLOW_USE_EIGEN_THREADPOOL is defined. This will allow to switch to the new non-blocking ThreadPool. Change: 119512280
* Fix RNN performance bug. + Additions to rnn benchmarks & benchmarks.py.Gravatar Eugene Brevdo2016-04-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The RNN performance bug: * When passing sequence_length to rnn(), calculations were being performed past max_sequence_length. This bug had one major side effect: * It slowed down the calculation past max_sequence_length (it *should* return zeros for outputs and copy state through) The calculations themselves were still correct: The state was still copied through and the output was still all zeros. But that calculation was performed via a vector-conditional select() instead of a single scalar cond(). As a result a lot of extra copying was happening both in fw and backprop. Thanks to Nat Roth (natusroth@gmail) for unearthing this bug. ************** Also: - updates to benchmarks.py (allow more specific benchmarks, added support for --benchmarks=all). - cleaned up RNN benchmarks code a bit. New and updated benchmarks: Calculation: Static Unroll with Halved Sequence Length vs. Half Static Unroll batch full_t units gpu dt(half_seq_len) dt(unroll_half) dt(half_seq_len)/dt(unroll_half) 128 50 256 False 0.164351 0.155019 1.060204 128 50 256 True 0.033295 0.028203 1.180550 Calculation: Static Unroll with Dynamic Flow LSTM vs. Dynamic Unroll LSTM batch max_t units gpu dt(static) dt(dynamic) dt(dynamic)/dt(static) 256 50 512 False 1.759111 1.692570 0.962173 256 50 512 True 0.178953 0.190454 1.064269 256 50 256 False 0.533132 0.567228 1.063955 256 50 256 True 0.078298 0.085024 1.085905 256 50 128 False 0.220362 0.215350 0.977255 256 50 128 True 0.053379 0.059129 1.107723 Change: 119495675
* Surface control_flow_ops.while_loop to public.Gravatar Yuan Yu2016-04-10
| | | | | Deprecated control_flow_ops.While. Use tf.while_loop. Change: 119488170
* This is another step to make TensorFlow more interactive and flexible to ↵Gravatar Yuan Yu2016-04-10
| | | | | | | | | | | | | | | | | | | | | | | | | | | users. It allows a tensor produced by a run call to stay "in-place" so that a future run call can use it in-place. To achieve this, a run call can now return a handle of a tensor to the client, which can then be fed to a subsequent run call. This feature is complimentary to partial run, though there are some overlaps. Here are a few properties of the current implementation: 1. Tensors are stored in the state of a session. The tensors are garbage collected if the client doesn't have a reference to the tensor or the session is closed. 2. There is no change to the current session API. We introduced two ops to manage the conversions between tensors and its handles. (There is a third op to garbage collect a tensor.) See the example below. 3. It fits quite well into the current feed-fetch design/implementation. It tries to reuse the graph (and caches) as much as possible so to make things efficient. Below is a simple example. More examples can be found in sessopn_ops_test.py. # Return a handle. a = tf.constant(10) b = tf.constant(5) c = tf.mul(a, b) h = tf.get_session_handle(c).eval() # Feed a tensor handle. f, x = tf.get_session_tensor(dtypes.int32) y = tf.mul(x, 10) result = sess.run(y, feed_dict={f: h.handle}) # result == 500 Change: 119481352
* Made isinf, isnan, isfinite, ceil and floor work with 16 bit floats.Gravatar Benoit Steiner2016-04-09
| | | | Change: 119458778
* Return correct status with error when Graph creation fails.Gravatar A. Unique TensorFlower2016-04-09
| | | | Change: 119448828
* Remove definition of macro TS_UNCHECKED_READ.Gravatar A. Unique TensorFlower2016-04-08
| | | | Change: 119434669
* Fix GPU BUILDGravatar Vijay Vasudevan2016-04-08
| | | | Change: 119431584
* Abstract the GCS/GFile code that event loading needs behind an io_wrapper ↵Gravatar A. Unique TensorFlower2016-04-08
| | | | | | | | | | | | | | module. io_wrapper provides some functions that check whether the path is a GCS path and calls the relevant functions from either gfile or gcs. This is *not* intended to be a general-purpose interface; it only implements the things that are necessary for loading events from GCS/GFile storage. We're doing this because having the entire event loader stack care about the difference between GCS and GFile is bad from an encapsulation perspective; this way, we can present one consistent interface. Change: 119427191
* Fix broken build.Gravatar A. Unique TensorFlower2016-04-08
| | | | Change: 119427077
* Temporary rollback of half support in TF_CALL_REAL_NUMBER_TYPES()Gravatar A. Unique TensorFlower2016-04-08
| | | | | due to internal issues. Change: 119424490
* Fix segfault in grpc_server_lib when an invalid hostname-port pair is passed.Gravatar Derek Murray2016-04-08
| | | | | | Previously, if the port was undefined, an out-of-bounds access would be made. This change adds the appropriate checks. Change: 119424297
* Use new adjoint attribute for solvers to make gradients more efficient.Gravatar A. Unique TensorFlower2016-04-08
| | | | | Consolidate linalg shape inference functions. Change: 119423897
* More comprehensively enforcing Tensor MaxDimensions limit.Gravatar David G. Andersen2016-04-08
| | | | Change: 119423048
* Remove unused mutex.h from file_system.hGravatar Manjunath Kudlur2016-04-08
| | | | Change: 119420831
* Added ParseExample microbenchmark.Gravatar Eugene Brevdo2016-04-08
| | | | Change: 119419160
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-06 00:55:03 +0000