From d57572e996dce24abf4d9cf6ea04e7104b3d743b Mon Sep 17 00:00:00 2001 From: Martin Wicke Date: Sat, 2 Sep 2017 19:21:45 -0700 Subject: Merge changes from github. PiperOrigin-RevId: 167401527 --- tensorflow/core/profiler/g3doc/advise.md | 4 ++-- tensorflow/core/profiler/g3doc/command_line.md | 6 +++--- tensorflow/core/profiler/g3doc/options.md | 8 ++++---- tensorflow/core/profiler/g3doc/profile_memory.md | 2 +- tensorflow/core/profiler/g3doc/profile_model_architecture.md | 8 ++++---- tensorflow/core/profiler/g3doc/profile_time.md | 12 ++++++------ 6 files changed, 20 insertions(+), 20 deletions(-) (limited to 'tensorflow/core/profiler/g3doc') diff --git a/tensorflow/core/profiler/g3doc/advise.md b/tensorflow/core/profiler/g3doc/advise.md index d87b0d8603..d0de8317f6 100644 --- a/tensorflow/core/profiler/g3doc/advise.md +++ b/tensorflow/core/profiler/g3doc/advise.md @@ -86,7 +86,7 @@ For example: * Checks RecvTensor RPC latency and bandwidth. * Checks CPU/Memory utilization of the job. -####AcceleratorUtilization Checker +#### AcceleratorUtilization Checker * Checks what percentage of time the accelerator spends on computation. #### OperationChecker @@ -100,7 +100,7 @@ For example: * Checks the most expensive graph nodes. * Checks the most expensive graph-building Python codes. -####Contribute Your Checker +#### Contribute Your Checker Follow examples of accelerator_utilization_checker.h diff --git a/tensorflow/core/profiler/g3doc/command_line.md b/tensorflow/core/profiler/g3doc/command_line.md index 857b5e6459..e2839a682f 100644 --- a/tensorflow/core/profiler/g3doc/command_line.md +++ b/tensorflow/core/profiler/g3doc/command_line.md @@ -51,7 +51,7 @@ It defines _checkpoint_variable op type. It also provides checkpointed tensors' Note: this feature is not well maintained now. -###Start `tfprof` +### Start `tfprof` #### Build `tfprof` @@ -140,9 +140,9 @@ tfprof> -output ``` -###Examples +### Examples -####Profile Python Time +#### Profile Python Time ```shell # Requires --graph_path --op_log_path tfprof> code -max_depth 1000 -show_name_regexes .*model_analyzer.*py.* -select micros -account_type_regexes .* -order_by micros diff --git a/tensorflow/core/profiler/g3doc/options.md b/tensorflow/core/profiler/g3doc/options.md index 15712d04c2..ddee63ad42 100644 --- a/tensorflow/core/profiler/g3doc/options.md +++ b/tensorflow/core/profiler/g3doc/options.md @@ -1,6 +1,6 @@ -##Options +## Options -###Overview +### Overview For all tfprof views, the profiles are processed with the following procedures @@ -35,7 +35,7 @@ For all tfprof views, the profiles are processed with the following procedures 4) Finally, the filtered data structure is output in a format depending on the `-output` option. -####Option Semantics In Different View +#### Option Semantics In Different View options usually have the same semantics in different views. However, some can vary. For example `-max_depth` in scope view means the depth of name scope tree. In op view, it means the length of operation list. @@ -68,7 +68,7 @@ output_bytes: The memory output by the operation. It's not necessarily requested by the current operation. For example, it can be a tensor forwarded from input to output, with in-place mutation. -###Docs +### Docs `-max_depth`: Show nodes that are at most this number of hops from starting node in the data structure. diff --git a/tensorflow/core/profiler/g3doc/profile_memory.md b/tensorflow/core/profiler/g3doc/profile_memory.md index a00683d062..6eda5abdd9 100644 --- a/tensorflow/core/profiler/g3doc/profile_memory.md +++ b/tensorflow/core/profiler/g3doc/profile_memory.md @@ -1,4 +1,4 @@ -##Profile Memory +## Profile Memory It is generally a good idea to visualize the memory usage in timeline. It allows you to see the memory consumption of each GPU over time. diff --git a/tensorflow/core/profiler/g3doc/profile_model_architecture.md b/tensorflow/core/profiler/g3doc/profile_model_architecture.md index a42b2e918d..61bb66bd21 100644 --- a/tensorflow/core/profiler/g3doc/profile_model_architecture.md +++ b/tensorflow/core/profiler/g3doc/profile_model_architecture.md @@ -1,9 +1,9 @@ -##Profile Model Architecture +## Profile Model Architecture * [Profile Model Parameters](#profile-model-parameters) * [Profile Model Float Operations](#profile-model-float-operations) -###Profile Model Parameters +### Profile Model Parameters Notes: `VariableV2` operation type might contain variables created by TensorFlow @@ -39,9 +39,9 @@ param_stats = tf.profiler.profile( sys.stdout.write('total_params: %d\n' % param_stats.total_parameters) ``` -###Profile Model Float Operations +### Profile Model Float Operations -####Caveats +#### Caveats For an operation to have float operation statistics: diff --git a/tensorflow/core/profiler/g3doc/profile_time.md b/tensorflow/core/profiler/g3doc/profile_time.md index e11a75553b..4aafc697a9 100644 --- a/tensorflow/core/profiler/g3doc/profile_time.md +++ b/tensorflow/core/profiler/g3doc/profile_time.md @@ -1,4 +1,4 @@ -##Profile Time +## Profile Time * [Times in TensorFlow and tfprof](#times-in-tensorflow-and-tfprof) * [Profile by Python Code](#profile-by-python-code) @@ -7,7 +7,7 @@ * [Profile by Name Scope](#profile-by-name-scope) -###Times in TensorFlow and tfprof +### Times in TensorFlow and tfprof When we run a model, Tensorflow schedules and runs the nodes (operations) in the graph. An operation can be placed on an accelerator or on CPU. @@ -37,7 +37,7 @@ When an operation is placed on CPU, it will completely run on CPU. Hence, should be 0. -###Profile by Python Code +### Profile by Python Code ```python # In code view, the time of each line of Python code is the aggregated # times of all operations created by that line. @@ -112,7 +112,7 @@ Set ```-output timeline:outfile=``` to generate timeline instead of st -###Profile by Operation Type +### Profile by Operation Type ```python # In op view, you can view the aggregated time of each operation type. tfprof> op -select micros,occurrence -order_by micros @@ -138,7 +138,7 @@ MatMul 618.97ms (63.56%, 16.51%), |/job:worker/replica:0/ ``` -###Profile by Graph +### Profile by Graph Usually, use graph view to generate a timeline to visualize the result. @@ -163,7 +163,7 @@ Open a Chrome browser, enter URL chrome://tracing and load the timeline file. ****************************************************** ``` -###Profile by Name Scope +### Profile by Name Scope Usually scope view allows you to pin point the problematic places if you have properly named your operations with tf.name_scope or tf.variable_scope. -- cgit v1.2.3