| /hardware/interfaces/neuralnetworks/utils/common/src/ |
| D | ResilientExecution.cpp | 42 auto execution = resilientExecution.getExecution(); in protect() local
43 auto result = fn(*execution); in protect()
51 auto maybeExecution = resilientExecution.recover(execution.get()); in protect()
59 execution = std::move(maybeExecution).value(); in protect()
61 return fn(*execution); in protect()
72 auto execution = NN_TRY(makeExecution()); in create() local
73 CHECK(execution != nullptr); in create()
75 std::move(execution)); in create()
79 nn::SharedExecution execution) in ResilientExecution() argument
80 : kMakeExecution(std::move(makeExecution)), mExecution(std::move(execution)) { in ResilientExecution()
[all …]
|
| /hardware/interfaces/neuralnetworks/utils/common/test/ |
| D | ResilientExecution.cpp | 91 const auto [mockExecution, mockExecutionFactory, execution] = setup(); in TEST()
94 const auto result = execution->getExecution(); in TEST()
102 const auto [mockExecution, mockExecutionFactory, execution] = setup(); in TEST()
106 const auto result = execution->compute({}); in TEST()
115 const auto [mockExecution, mockExecutionFactory, execution] = setup(); in TEST()
119 const auto result = execution->compute({}); in TEST()
128 const auto [mockExecution, mockExecutionFactory, execution] = setup(); in TEST()
133 const auto result = execution->compute({}); in TEST()
142 const auto [mockExecution, mockExecutionFactory, execution] = setup(); in TEST()
149 const auto result = execution->compute({}); in TEST()
[all …]
|
| /hardware/interfaces/neuralnetworks/1.2/ |
| D | IPreparedModel.hal | 27 * IPreparedModel describes a model that has been prepared for execution and
32 * Launches an asynchronous execution on a prepared model.
34 * The execution is performed asynchronously with respect to the caller.
39 * launch an asynchronous task to perform the execution in the background,
45 * When the asynchronous task has finished its execution, it must
48 * the execution.
60 * - the execution should launch successfully (ErrorStatus::NONE): There
62 * - if at execution time every operation's input operands have legal
63 * values, the execution should complete successfully (ErrorStatus::NONE):
72 * @param measure Specifies whether or not to measure duration of the execution.
[all …]
|
| D | IExecutionCallback.hal | 24 * execution asynchronously launched from IPreparedModel::execute.
30 * task has finished performing the execution. Either notify_1_2 or notify must be
31 * provided with the ErrorStatus from the execution. If the asynchronous task is
38 * - NONE if the asynchronous execution was successful
52 * @return Timing Duration of execution. Unless MeasureTiming::YES was passed when
53 * launching the execution and status is NONE, all times must
|
| D | types.t | 120 * execution request (a {@link @1.0::Request} object and a {@link MeasureTiming}
123 * The serialized representation for a particular execution is referred to later
223 * for the inputs and outputs of an execution. However, these memory pools
237 * Specifies whether or not to measure duration of the execution. The
247 * values returned from an execution ({@link @1.0::ErrorStatus},
251 * The serialized representation for a particular execution is referred to later
257 * The execution return values ({@link @1.0::ErrorStatus} and
278 * Status of the execution.
307 * execution.
322 * Duration of execution. Unless measurement was requested and execution
|
| D | IDevice.hal | 134 * have no worse effect than generating bad output values at execution time.
138 * to the model cache may affect the driver's execution behavior, and a malicious
168 * Asynchronously creates a prepared model for execution and optionally saves it
172 * representations to a model for execution, possibly including
222 * the shape of the tensors, which may only be known at execution time. As
231 * @param model The model to be prepared for execution.
232 * @param preference Indicates the intended execution behavior of a prepared
251 * in a failed execution or in a successful execution that produces incorrect
255 * preparing the model for execution and the prepared model if
273 * Creates a prepared model from cache files for execution.
[all …]
|
| D | IPreparedModelCallback.hal | 50 * execution. If the model was unable to be prepared
|
| D | types.hal | 2672 * This op applies NMS algorithm to each class. In each loop of execution,
4923 * This includes not only the execution graph, but also constant data such as
4936 * The operations are sorted into execution order. Every operand
5039 * Describes the shape information of an output operand after execution.
5054 * Specifies whether or not to measure timing information during execution.
5063 * Timing information measured during execution. Each time is a duration from
5080 * execution request (a {@link @1.0::Request} object and a {@link MeasureTiming}
5083 * The serialized representation for a particular execution is referred to later
5183 * for the inputs and outputs of an execution. However, these memory pools
5197 * Specifies whether or not to measure duration of the execution. The
[all …]
|
| /hardware/interfaces/neuralnetworks/1.0/ |
| D | IPreparedModel.hal | 22 * IPreparedModel describes a model that has been prepared for execution and
27 * Launches an asynchronous execution on a prepared model.
29 * The execution is performed asynchronously with respect to the caller.
34 * launch an asynchronous task to perform the execution in the background,
40 * When the asynchronous task has finished its execution, it must
43 * the execution.
55 * - the execution should launch successfully (ErrorStatus::NONE): There
57 * - if at execution time every operation's input operands have legal
58 * values, the execution should complete successfully (ErrorStatus::NONE):
67 * the execution. The callback object's notify function must
[all …]
|
| D | IExecutionCallback.hal | 21 * execution asynchronously launched from IPreparedModel::execute.
27 * finished performing the execution. notify must be provided with the
28 * ErrorStatus resulting from the execution. If the asynchronous task
34 * - NONE if the asynchronous execution was successful
|
| D | IDevice.hal | 61 * Creates a prepared model for execution.
64 * representations to a model for execution, possiblly including
90 * the shape of the tensors, which may only be known at execution time. As
99 * @param model The model to be prepared for execution.
101 * preparing the model for execution and the prepared model
|
| D | IPreparedModelCallback.hal | 48 * execution. If the model was unable to be prepared
|
| D | types.hal | 1707 * This includes not only the execution graph, but also constant data such as
1720 * The operations are sorted into execution order. Every operand
1807 * Input data and information to be used in the execution of a prepared
1816 * Output data and information to be used in the execution of a prepared
|
| /hardware/interfaces/neuralnetworks/1.3/ |
| D | IExecutionCallback.hal | 25 * execution asynchronously launched from IPreparedModel::execute*.
33 * finished performing the execution. One of the notify methods must be
34 * provided with the ErrorStatus from the execution. If the asynchronous
41 * - NONE if the asynchronous execution was successful
50 * - MISSED_DEADLINE_* if the execution is aborted because it
58 * @param timing Duration of execution. Unless MeasureTiming::YES was passed when
59 * launching the execution and status is NONE, all times must
|
| D | IFencedExecutionCallback.hal | 35 * @return status Error status returned from the asynchronously dispatched execution
37 * - NONE if the asynchronous execution was successful
41 * - MISSED_DEADLINE_* if the execution is aborted because it
48 * launching the execution and status is NONE, all times
55 * launching the execution and status is NONE, all times
|
| D | IDevice.hal | 78 * Asynchronously creates a prepared model for execution and optionally
82 * alternative representations to a model for execution, possibly including
152 * the shape of the tensors, which may only be known at execution time. As
161 * @param model The model to be prepared for execution.
162 * @param preference Indicates the intended execution behavior of a prepared
188 * collision; a collision will result in a failed execution or in a
189 * successful execution that produces incorrect output values. If both
193 * preparing the model for execution and the prepared model if
215 * Creates a prepared model from cache files for execution.
271 * the shape of the tensors, which may only be known at execution time. As
[all …]
|
| D | types.t | 174 * execution timeout duration associated with it ("loop timeout duration").
175 * This duration is configurable on a per-execution basis and must not exceed
|
| D | IPreparedModelCallback.hal | 55 * execution. If the model was unable to be prepared
|
| D | types.hal | 2833 * This op applies NMS algorithm to each class. In each loop of execution,
5124 * * 3 ~ (n + 2): Inputs to be passed to the subgraph selected for execution.
5335 * Priority given to a prepared model for execution.
5596 * This includes not only the execution graph, but also constant data such as
5667 * An excerpt of the execution graph.
5678 * The operations are sorted into execution order. Every operand
5748 * Input data and information to be used in the execution of a prepared
5757 * Output data and information to be used in the execution of a prepared
5849 * execution timeout duration associated with it ("loop timeout duration").
5850 * This duration is configurable on a per-execution basis and must not exceed
|
| /hardware/interfaces/neuralnetworks/1.1/ |
| D | IDevice.hal | 66 * Creates a prepared model for execution.
69 * representations to a model for execution, possiblly including
95 * the shape of the tensors, which may only be known at execution time. As
104 * @param model The model to be prepared for execution.
105 * @param preference Indicates the intended execution behavior of a prepared
108 * preparing the model for execution and the prepared model
|
| D | types.hal | 403 * This includes not only the execution graph, but also constant data such as
416 * The operations are sorted into execution order. Every operand
|
| /hardware/interfaces/neuralnetworks/utils/common/include/nnapi/hal/ |
| D | ResilientExecution.h | 44 nn::SharedExecution execution);
|
| /hardware/interfaces/dumpstate/1.1/ |
| D | IDumpstateDevice.hal | 40 * If execution runs longer than this, the IDumpstateDevice service may be killed and only
|
| /hardware/interfaces/audio/policy/1.0/xml/pfw_schemas/ |
| D | README.md | 28 - a relative path (relative to the execution directory),
|
| /hardware/interfaces/wifi/supplicant/1.0/ |
| D | ISupplicantStaIface.hal | 421 * execution. Timeout must be set to |ExtRadioWorkDefaults.TIMEOUT_IN_SECS|
|