B ӻd@@sDdZddlZddlmZddlmZedgdGdddeZdS) z(Utilities for building profiler options.N) tfprof_logger) tf_exportzprofiler.ProfileOptionBuilder)Zv1c@seZdZdZd1ddZeddZeddZed2d d Zd dZ ddZ d3ddZ d4ddZ ddZ ddZddZddZd5ddZdd Zd!d"Zd#d$Zd%d&Zd'd(Zd)d*Zd+d,Zd-d.Zd/d0ZdS)6ProfileOptionBuilderaOption Builder for Profiling API. For tutorial on the options, see https://github.com/tensorflow/tensorflow/tree/master/tensorflow/core/profiler/g3doc/options.md ```python # Users can use pre-built options: opts = ( tf.profiler.ProfileOptionBuilder.trainable_variables_parameter()) # Or, build your own options: opts = (tf.compat.v1.profiler.ProfileOptionBuilder() .with_max_depth(10) .with_min_micros(1000) .select(['accelerator_micros']) .with_stdout_output() .build() # Or customize the pre-built options: opts = (tf.compat.v1.profiler.ProfileOptionBuilder( tf.profiler.ProfileOptionBuilder.time_and_memory()) .with_displaying_options(show_name_regexes=['.*rnn.*']) .build()) # Finally, profiling with the options: _ = tf.compat.v1.profiler.profile(tf.compat.v1.get_default_graph(), run_meta=run_meta, cmd='scope', options=opts) ``` NcCsJ|dk rt||_n0ddddddddgdggdggddgdd d |_dS) zWConstructor. Args: options: Optional initial option dict to start with. Ndrnamez.*Fmicrosstdout) max_depth min_bytes min_micros min_params min_float_opsmin_occurrenceorder_byaccount_type_regexesstart_name_regexestrim_name_regexesshow_name_regexeshide_name_regexesaccount_displayed_op_onlyselectstepoutput)copydeepcopy_options)selfoptionsr[/var/www/html/venv/lib/python3.7/site-packages/tensorflow/python/profiler/option_builder.py__init__:s$zProfileOptionBuilder.__init__cCs0dddddddtjgdggdggddgddd S) zOptions used to profile trainable variable parameters. Normally used together with 'scope' view. Returns: A dict of profiling options. i'rrz.*Tparamsrr )r r r r rrrrrrrrrrrr)rZTRAINABLE_VARIABLESrrrr trainable_variables_parameterTs z2ProfileOptionBuilder.trainable_variables_parametercCs.ddddddddgdggdggddgddd S) aOptions used to profile float operations. Please see https://github.com/tensorflow/tensorflow/tree/master/tensorflow/core/profiler/g3doc/profile_model_architecture.md on the caveats of calculating float operations. Returns: A dict of profiling options. i'rZ float_opsz.*Trr )r r r r rrrrrrrrrrrrrrrrr float_operationns z$ProfileOptionBuilder.float_operationr$rcCs:d|||||||dddddgdggdggdddgddd S) aShow operation time and memory consumptions. Args: min_micros: Only show profiler nodes with execution time no less than this. It sums accelerator and cpu times. min_bytes: Only show profiler nodes requested to allocate no less bytes than this. min_accelerator_micros: Only show profiler nodes spend no less than this time on accelerator (e.g. GPU). min_cpu_micros: Only show profiler nodes spend no less than this time on cpu. min_peak_bytes: Only show profiler nodes using no less than this bytes at peak (high watermark). For profiler nodes consist of multiple graph nodes, it sums the graph nodes' peak_bytes. min_residual_bytes: Only show profiler nodes have no less than this bytes not being de-allocated after Compute() ends. For profiler nodes consist of multiple graph nodes, it sums the graph nodes' residual_bytes. min_output_bytes: Only show profiler nodes have no less than this bytes output. The output are not necessarily allocated by this profiler nodes. Returns: A dict of profiling options. i'rrz.*Tbytesrr )r r min_peak_bytesmin_residual_bytesmin_output_bytesr min_accelerator_microsmin_cpu_microsr rrrrrrrrrrrrr)r r r*r+r'r(r)rrr time_and_memorys*z$ProfileOptionBuilder.time_and_memorycCs t|jS)zOBuild a profiling option. Returns: A dict of profiling options. )rrr)rrrr buildszProfileOptionBuilder.buildcCs||jd<|S)a<Set the maximum depth of display. The depth depends on profiling view. For 'scope' view, it's the depth of name scope hierarchy (tree), for 'op' view, it's the number of operation types (list), etc. Args: max_depth: Maximum depth of the data structure to display. Returns: self r )r)rr rrr with_max_depths z#ProfileOptionBuilder.with_max_depthcCs,||jd<||jd<||jd<||jd<|S)a>Only show profiler nodes consuming no less than 'min_bytes'. Args: min_bytes: Only show profiler nodes requested to allocate no less bytes than this. min_peak_bytes: Only show profiler nodes using no less than this bytes at peak (high watermark). For profiler nodes consist of multiple graph nodes, it sums the graph nodes' peak_bytes. min_residual_bytes: Only show profiler nodes have no less than this bytes not being de-allocated after Compute() ends. For profiler nodes consist of multiple graph nodes, it sums the graph nodes' residual_bytes. min_output_bytes: Only show profiler nodes have no less than this bytes output. The output are not necessarily allocated by this profiler nodes. Returns: self r r'r(r))r)rr r'r(r)rrr with_min_memorys     z$ProfileOptionBuilder.with_min_memorycCs"||jd<||jd<||jd<|S)aOnly show profiler nodes consuming no less than 'min_micros'. Args: min_micros: Only show profiler nodes with execution time no less than this. It sums accelerator and cpu times. min_accelerator_micros: Only show profiler nodes spend no less than this time on accelerator (e.g. GPU). min_cpu_micros: Only show profiler nodes spend no less than this time on cpu. Returns: self r r*r+)r)rr r*r+rrr with_min_execution_times   z,ProfileOptionBuilder.with_min_execution_timecCs||jd<|S)aDOnly show profiler nodes holding no less than 'min_params' parameters. 'Parameters' normally refers the weights of in TensorFlow variables. It reflects the 'capacity' of models. Args: min_params: Only show profiler nodes holding number parameters no less than this. Returns: self r )r)rr rrr with_min_parameterss z(ProfileOptionBuilder.with_min_parameterscCs||jd<|S)aOnly show profiler nodes including no less than 'min_occurrence' graph nodes. A "node" means a profiler output node, which can be a python line (code view), an operation type (op view), or a graph node (graph/scope view). A python line includes all graph nodes created by that line, while an operation type includes all graph nodes of that type. Args: min_occurrence: Only show nodes including no less than this. Returns: self r)r)rrrrr with_min_occurrences z(ProfileOptionBuilder.with_min_occurrencecCs||jd<|S)aOnly show profiler nodes consuming no less than 'min_float_ops'. Please see https://github.com/tensorflow/tensorflow/tree/master/tensorflow/core/profiler/g3doc/profile_model_architecture.md on the caveats of calculating float operations. Args: min_float_ops: Only show profiler nodes with float operations no less than this. Returns: self r)r)rrrrr with_min_float_operations's z.ProfileOptionBuilder.with_min_float_operationscCst||jd<|S)aSelectively counting statistics based on node types. Here, 'types' means the profiler nodes' properties. Profiler by default consider device name (e.g. /job:xx/.../device:GPU:0) and operation type (e.g. MatMul) as profiler nodes' properties. User can also associate customized 'types' to profiler nodes through OpLogProto proto. For example, user can select profiler nodes placed on gpu:0 with: `account_type_regexes=['.*gpu:0.*']` If none of a node's properties match the specified regexes, the node is not displayed nor accounted. Args: account_type_regexes: A list of regexes specifying the types. Returns: self. r)rr)rrrrr with_accounted_types8sz)ProfileOptionBuilder.with_accounted_typescCsd|dk rt||jd<|dk r0t||jd<|dk rHt||jd<|dk r`t||jd<|S)aMRegular expressions used to select profiler nodes to display. After 'with_accounted_types' is evaluated, 'with_node_names' are evaluated as follows: For a profile data structure, profiler first finds the profiler nodes matching 'start_name_regexes', and starts displaying profiler nodes from there. Then, if a node matches 'show_name_regexes' and doesn't match 'hide_name_regexes', it's displayed. If a node matches 'trim_name_regexes', profiler stops further searching that branch. Args: start_name_regexes: list of node name regexes to start displaying. show_name_regexes: list of node names regexes to display. hide_name_regexes: list of node_names regexes that should be hidden. trim_name_regexes: list of node name regexes from where to stop. Returns: self Nrrrr)rr)rrrrrrrr with_node_namesNsz$ProfileOptionBuilder.with_node_namescCs||jd<|S)aWhether only account the statistics of displayed profiler nodes. Args: is_true: If true, only account statistics of nodes eventually displayed by the outputs. Otherwise, a node's statistics are accounted by its parents as long as it's types match 'account_type_regexes', even if it is hidden from the output, say, by hide_name_regexes. Returns: self r)r)rZis_truerrr rps z.ProfileOptionBuilder.account_displayed_op_onlycCsd|jd<|S)z$Do not generate side-effect outputs.noner)r)rrrr with_empty_outputs z&ProfileOptionBuilder.with_empty_outputcCsd|jd<|S)zPrint the result to stdout.r r)r)rrrr with_stdout_outputs z'ProfileOptionBuilder.with_stdout_outputcCsd||jd<|S)zPrint the result to a file.zfile:outfile=%sr)r)routfilerrr with_file_outputsz%ProfileOptionBuilder.with_file_outputcCsd||jd<|S)zGenerate a timeline json file.ztimeline:outfile=%sr)r)rZ timeline_filerrr with_timeline_outputsz)ProfileOptionBuilder.with_timeline_outputcCsd||jd<|S)zGenerate a pprof profile gzip file. To use the pprof file: pprof -png --nodecount=100 --sample_index=1 Args: pprof_file: filename for output, usually suffixed with .pb.gz. Returns: self. zpprof:outfile=%sr)r)rZ pprof_filerrr with_pprof_outputs z&ProfileOptionBuilder.with_pprof_outputcCs||jd<|S)aBOrder the displayed profiler nodes based on a attribute. Supported attribute includes micros, bytes, occurrence, params, etc. https://github.com/tensorflow/tensorflow/tree/master/tensorflow/core/profiler/g3doc/options.md Args: attribute: An attribute the profiler node has. Returns: self r)r)r attributerrr rs zProfileOptionBuilder.order_bycCst||jd<|S)a Select the attributes to display. See https://github.com/tensorflow/tensorflow/tree/master/tensorflow/core/profiler/g3doc/options.md for supported attributes. Args: attributes: A list of attribute the profiler node has. Returns: self r)rr)r attributesrrr rs zProfileOptionBuilder.selectcCs||jd<|S)a1Which profile step to use for profiling. The 'step' here refers to the step defined by `Profiler.add_step()` API. Args: step: When multiple steps of profiles are available, select which step's profile to use. If -1, use average of all available steps. Returns: self r)r)rrrrr with_steps zProfileOptionBuilder.with_step)N)r$r$rrrrr)rrrr)rrr)NNNN)__name__ __module__ __qualname____doc__r! staticmethodr#r%r,r-r.r/r0r1r2r3r4r5rr7r8r:r;r<rrr?rrrr rsD!    /   r)rCrZtensorflow.python.profilerrZ tensorflow.python.util.tf_exportrobjectrrrrr s