Commit Graph

23 Commits

Author SHA1 Message Date
Yasir Modak
1043101a8b
updating dead web link
fixes 37047
2020-02-26 15:35:05 -08:00
Kazuaki Ishizaki
bd8e308b4c minor spelling tweaks 2020-02-11 15:09:21 +09:00
Hye Soo Yang
ffba3db281 Automated rollback of commit 97b65731ac
PiperOrigin-RevId: 268547754
2019-09-11 15:04:59 -07:00
Hye Soo Yang
97b65731ac Automated rollback of commit b2c58f3225
PiperOrigin-RevId: 268523903
2019-09-11 13:17:50 -07:00
Hye Soo Yang
b2c58f3225 Export the tfprof functions from C++ to Python with pybind11 instead of swig. This is part of a larger effort to deprecate swig and eventually with modularization break pywrap_tensorflow into smaller components. It will also make exporting C++ ops to Python significantly easier. XLA is using the pybind11 macros already. Please refer to https://github.com/tensorflow/community/blob/master/rfcs/20190208-pybind11.md for more information.
PiperOrigin-RevId: 267530586
2019-09-05 22:56:09 -07:00
Mark Daoust
1bde2c4ce9 Apply tf1->tf2 name replaces to doc-strings and comments in tensorflow.
No code changes, only doc-strings and comments (I checked).

PiperOrigin-RevId: 243885306
2019-04-16 16:27:28 -07:00
Martin Wicke
ac79b5c1d2 Remove symbols from the v2 API. Included here:
tf.train.NewCheckpointReader
tf.train.replica_device_setter
tf.train.init_from_checkpoint
tf.train.ProfilerHook
tf.profiler.*

Profiler's API depends on v1 style graphs, so we'll remove this pending replacement.

PiperOrigin-RevId: 221359471
2018-11-13 17:00:13 -08:00
gehring
c9b4689bc4
Removed no longer supported call to in_eager_execution
changes to model_analyser.analyse(...):
- Swapped context.in_eager_execution() to the currently supported context.executing_eagerly().
- Added negation to eager check. In all likelihood, the negation was always supposed to be there since getting default graph in eager mode does not make sense. The current `if` condition is likely a bug. The proposed fix is consistent with other functions in this module, e.g., `profile(...)`, line 339.
2018-11-06 12:32:01 -05:00
Asim Shankar
37cef895bf eager: Rename in_eager_mode to executing_eagerly and get rid of in_graph_mode.
This is in preparation to introduce one public, stable symbol: tf.executing_eagerly()
(i.e., part of moving APIs related to eager execution from "contrib" to a namespace
where we provide API stability guarantees)

PiperOrigin-RevId: 188212646
2018-03-07 12:10:42 -08:00
Anna R
adc9ee7150 Adding tf_export decorators/calls to TensorFlow functions and constants.
PiperOrigin-RevId: 184020524
2018-01-31 17:20:42 -08:00
A. Unique TensorFlower
5a814028e2 tfprof support for eager.
PiperOrigin-RevId: 181784078
2018-01-12 13:24:39 -08:00
A. Unique TensorFlower
118495de61 profiler C++ API.
PiperOrigin-RevId: 181397308
2018-01-09 17:13:03 -08:00
Shanqing Cai
93dc52c707 tfdbg: Add protocol for sending tracebacks and source code to debug server
PiperOrigin-RevId: 178192708
2017-12-06 20:39:10 -08:00
A. Unique TensorFlower
1bc367859c A few profiler improvements.
1. Op view proto copy uses too much memory and time, optimized.
2. Add a hint to use "bazel-bin" instead of "bazel run"
3. Make proto string parsing explicit (seems no longer throwing error)

PiperOrigin-RevId: 175745677
2017-11-14 15:37:15 -08:00
A. Unique TensorFlower
113be57466 A few profiler improvements
1. Track the full allocation history of each tensor, visualized in timeline.
2. Better ProfileContext for tracing step selection.
3. Small bug fix.

PiperOrigin-RevId: 173988293
2017-10-30 20:13:00 -07:00
A. Unique TensorFlower
b76565b39d Some profiler fixes and cleanup.
PiperOrigin-RevId: 168069346
2017-09-08 17:34:54 -07:00
A. Unique TensorFlower
9e9ffa33d1 Unify all profile files (graph,run_meta,op_log) into one.
Also allow profiler to serialize/deserialize to/from file.

PiperOrigin-RevId: 167815923
2017-09-06 22:01:54 -07:00
A. Unique TensorFlower
e722358e7e Merge changes from github.
END_PUBLIC

---
Commit 607816029 authored by Eugene Brevdo<ebrevdo@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Extended ScratchSpace to expose its underlying scratch tensor object.

PiperOrigin-RevId: 167649551

---
Commit db43fe68e authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add fast math attributes to all generated methods when fast math enabled.

RELNOTES: n/a
PiperOrigin-RevId: 167646637

---
Commit aebe8cc6f authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Call HloComputation.Accept instead of HloInstruction.Accept to get all instructions profiled.

RELNOTES: n/a
PiperOrigin-RevId: 167640259

---
Commit 0ab137cd8 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
BEGIN_PUBLIC
Automated g4 rollback of changelist 167604306

PiperOrigin-RevId: 167800256
2017-09-06 18:01:27 -07:00
A. Unique TensorFlower
2b51e0ba27 1. Support profiling nodes in RunMetadata but not in GraphDef.
2. Add an API to allow easier profile retrieval. Currently in contrib.

PiperOrigin-RevId: 165399640
2017-08-15 21:42:58 -07:00
A. Unique TensorFlower
19c27ef0d5 Fine-grained memory profiling
Add residual_bytes, peak_bytes and output_bytes.
Allow to order/select/filter by
accelerator_micros/cpu_micros/peak_bytes/residual_bytes/output_bytes

Also updated the testdata.

PiperOrigin-RevId: 164079214
2017-08-02 21:34:37 -07:00
A. Unique TensorFlower
8281e234c1 tfprof further clean ups
PiperOrigin-RevId: 161486865
2017-07-10 22:16:36 -07:00
A. Unique TensorFlower
154df32a95 Follow up changes of tfprof migration:
1. Add a option_builder module to build options for Python API.
2. Expose necessary profiler protos.
3. Rename command line tool from tfprof to profiler.

PiperOrigin-RevId: 161160274
2017-07-06 19:10:31 -07:00
A. Unique TensorFlower
af23ae65db Migrating tfprof python API to tensorflow/python/profiler
Migrating tfprof c++ to tensorflow/core/profiler

API changes:
New tf.profiler namespace.

Within tf.profiler namespace:
tf.profiler.advise  # One-shot advise function
tf.profiler.profile  # One-shot profile function
tf.profiler.Profiler  # Multi-step profile/advise class
tf.profiler.write_op_log  # Write profile for offline analysis

PiperOrigin-RevId: 160901831
2017-07-04 10:23:59 -07:00