LostTech.TensorFlow : API Documentation

Type Benchmark

Namespace tensorflow.test

Parent Benchmark

Interfaces IBenchmark

Abstract class that provides helpers for TensorFlow benchmarks.

Methods

Properties

Public instance methods

object evaluate(PythonClassContainer tensors)

Evaluates tensors and returns numpy values.
Parameters
PythonClassContainer tensors
A Tensor or a nested list/tuple of Tensors.
Returns
object
tensors numpy values.

object evaluate(IGraphNodeBase tensors)

Evaluates tensors and returns numpy values.
Parameters
IGraphNodeBase tensors
A Tensor or a nested list/tuple of Tensors.
Returns
object
tensors numpy values.

object evaluate(IEnumerable<object> tensors)

Evaluates tensors and returns numpy values.
Parameters
IEnumerable<object> tensors
A Tensor or a nested list/tuple of Tensors.
Returns
object
tensors numpy values.

object evaluate_dyn(object tensors)

Evaluates tensors and returns numpy values.
Parameters
object tensors
A Tensor or a nested list/tuple of Tensors.
Returns
object
tensors numpy values.

void report_benchmark(Nullable<int> iters, object cpu_time, double wall_time, Nullable<double> throughput, IDictionary<string, object> extras, string name, IEnumerable<IDictionary<string, object>> metrics)

void report_benchmark(Nullable<int> iters, object cpu_time, Nullable<int> wall_time, Nullable<double> throughput, IDictionary<string, object> extras, string name, IEnumerable<IDictionary<string, object>> metrics)

object report_benchmark_dyn(object iters, object cpu_time, object wall_time, object throughput, object extras, object name, object metrics)

IDictionary<string, object> run_op_benchmark(Session sess, IGraphNodeBase op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, string name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IGraphNodeBase op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
string name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IGraphNodeBase op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, int, object> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IGraphNodeBase op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, int, object> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, object op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, int, object> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
object op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, int, object> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, object op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, object> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
object op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, object> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, object op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, string name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
object op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
string name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IGraphNodeBase op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, int, object, object, object, int> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IGraphNodeBase op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, int, object, object, object, int> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IEnumerable<object> op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, string name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IEnumerable<object> op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
string name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IEnumerable<object> op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, object> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IEnumerable<object> op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, object> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IEnumerable<object> op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, int, object> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IEnumerable<object> op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, int, object> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IEnumerable<object> op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, int, object, object, object, int> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IEnumerable<object> op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, int, object, object, object, int> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, IGraphNodeBase op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, object> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
IGraphNodeBase op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, object> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

IDictionary<string, object> run_op_benchmark(Session sess, object op_or_tensor, IDictionary<object, object> feed_dict, int burn_iters, int min_iters, bool store_trace, bool store_memory_usage, ValueTuple<int, int, object, object, object, int> name, IDictionary<string, object> extras, int mbs)

Run an op or tensor in the given session. Report the results.
Parameters
Session sess
`Session` object to use for timing.
object op_or_tensor
`Operation` or `Tensor` to benchmark.
IDictionary<object, object> feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
int burn_iters
Number of burn-in iterations to run.
int min_iters
Minimum number of iterations to use for timing.
bool store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
bool store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
ValueTuple<int, int, object, object, object, int> name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
IDictionary<string, object> extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
int mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
IDictionary<string, object>
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

object run_op_benchmark_dyn(object sess, object op_or_tensor, object feed_dict, ImplicitContainer<T> burn_iters, ImplicitContainer<T> min_iters, ImplicitContainer<T> store_trace, ImplicitContainer<T> store_memory_usage, object name, object extras, ImplicitContainer<T> mbs)

Run an op or tensor in the given session. Report the results.
Parameters
object sess
`Session` object to use for timing.
object op_or_tensor
`Operation` or `Tensor` to benchmark.
object feed_dict
A `dict` of values to feed for each op iteration (see the `feed_dict` parameter of `Session.run`).
ImplicitContainer<T> burn_iters
Number of burn-in iterations to run.
ImplicitContainer<T> min_iters
Minimum number of iterations to use for timing.
ImplicitContainer<T> store_trace
Boolean, whether to run an extra untimed iteration and store the trace of iteration in returned extras. The trace will be stored as a string in Google Chrome trace format in the extras field "full_trace_chrome_format". Note that trace will not be stored in test_log_pb2.TestResults proto.
ImplicitContainer<T> store_memory_usage
Boolean, whether to run an extra untimed iteration, calculate memory usage, and store that in extras fields.
object name
(optional) Override the BenchmarkEntry name with `name`. Otherwise it is inferred from the top-level method name.
object extras
(optional) Dict mapping string keys to additional benchmark info. Values may be either floats or values that are convertible to strings.
ImplicitContainer<T> mbs
(optional) The number of megabytes moved by this op, used to calculate the ops throughput.
Returns
object
A `dict` containing the key-value pairs that were passed to `report_benchmark`. If `store_trace` option is used, then `full_chrome_trace_format` will be included in return dictionary even though it is not passed to `report_benchmark` with `extras`.

Public static methods

object is_abstract_dyn<TClass>()

TClass is_abstract<TClass>()

Public properties

object PythonObject get;