Go to the documentation of this file.
176 static constexpr int32_t kVALUE = 12;
210 mImpl->setName(name);
222 return mImpl->getName();
241 mImpl->setDimensions(dimensions);
254 return mImpl->getDimensions();
269 mImpl->setType(type);
281 return mImpl->getType();
296 return mImpl->setDynamicRange(min, max);
304 return mImpl->isNetworkInput();
312 return mImpl->isNetworkOutput();
334 mImpl->setBroadcastAcrossBatch(broadcastAcrossBatch);
350 return mImpl->getBroadcastAcrossBatch();
360 return mImpl->getLocation();
375 mImpl->setLocation(location);
385 return mImpl->dynamicRangeIsSet();
393 mImpl->resetDynamicRange();
403 return mImpl->getDynamicRangeMin();
413 return mImpl->getDynamicRangeMax();
432 mImpl->setAllowedFormats(formats);
445 return mImpl->getAllowedFormats();
479 return mImpl->isShapeTensor();
502 return mImpl->isExecutionTensor();
506 apiv::VTensor* mImpl;
507 virtual ~
ITensor() noexcept = default;
527 return mLayer->getType();
539 mLayer->setName(name);
550 return mLayer->getName();
558 return mLayer->getNbInputs();
571 return mLayer->getInput(index);
579 return mLayer->getNbOutputs();
590 return mLayer->getOutput(index);
607 return mLayer->setInput(index, tensor);
627 mLayer->setPrecision(dataType);
639 return mLayer->getPrecision();
651 return mLayer->precisionIsSet();
661 mLayer->resetPrecision();
692 mLayer->setOutputType(index, dataType);
706 return mLayer->getOutputType(index);
719 return mLayer->outputTypeIsSet(index);
731 return mLayer->resetOutputType(index);
735 virtual ~
ILayer() noexcept = default;
736 apiv::VLayer* mLayer;
977 static constexpr int32_t kVALUE = 6;
1007 mImpl->setKernelSize(kernelSize);
1019 return mImpl->getKernelSize();
1031 mImpl->setNbOutputMaps(nbOutputMaps);
1041 return mImpl->getNbOutputMaps();
1057 mImpl->setStride(stride);
1067 return mImpl->getStride();
1087 return mImpl->setPadding(padding);
1099 return mImpl->getPadding();
1119 mImpl->setNbGroups(nbGroups);
1129 return mImpl->getNbGroups();
1143 mImpl->setKernelWeights(weights);
1153 return mImpl->getKernelWeights();
1168 mImpl->setBiasWeights(weights);
1178 return mImpl->getBiasWeights();
1194 return mImpl->setDilation(dilation);
1206 return mImpl->getDilation();
1223 mImpl->setPrePadding(padding);
1233 return mImpl->getPrePadding();
1250 mImpl->setPostPadding(padding);
1260 return mImpl->getPostPadding();
1274 mImpl->setPaddingMode(paddingMode);
1286 return mImpl->getPaddingMode();
1299 mImpl->setKernelSizeNd(kernelSize);
1309 return mImpl->getKernelSizeNd();
1324 mImpl->setStrideNd(stride);
1334 return mImpl->getStrideNd();
1352 mImpl->setPaddingNd(padding);
1364 return mImpl->getPaddingNd();
1378 mImpl->setDilationNd(dilation);
1388 return mImpl->getDilationNd();
1416 apiv::VConvolutionLayer* mImpl;
1460 mImpl->setNbOutputChannels(nbOutputs);
1470 return mImpl->getNbOutputChannels();
1480 mImpl->setKernelWeights(weights);
1490 return mImpl->getKernelWeights();
1502 mImpl->setBiasWeights(weights);
1512 return mImpl->getBiasWeights();
1540 apiv::VFullyConnectedLayer* mImpl;
1566 mImpl->setActivationType(type);
1576 return mImpl->getActivationType();
1591 mImpl->setAlpha(alpha);
1605 mImpl->setBeta(beta);
1614 return mImpl->getAlpha();
1623 return mImpl->getBeta();
1628 apiv::VActivationLayer* mImpl;
1640 kMAX_AVERAGE_BLEND = 2
1649 static constexpr int32_t kVALUE = 3;
1676 mImpl->setPoolingType(type);
1686 return mImpl->getPoolingType();
1700 mImpl->setWindowSize(windowSize);
1712 return mImpl->getWindowSize();
1728 mImpl->setStride(stride);
1740 return mImpl->getStride();
1756 mImpl->setPadding(padding);
1770 return mImpl->getPadding();
1785 mImpl->setBlendFactor(blendFactor);
1798 return mImpl->getBlendFactor();
1815 mImpl->setAverageCountExcludesPadding(exclusive);
1826 return mImpl->getAverageCountExcludesPadding();
1844 mImpl->setPrePadding(padding);
1854 return mImpl->getPrePadding();
1872 mImpl->setPostPadding(padding);
1882 return mImpl->getPostPadding();
1895 mImpl->setPaddingMode(paddingMode);
1906 return mImpl->getPaddingMode();
1919 mImpl->setWindowSizeNd(windowSize);
1929 return mImpl->getWindowSizeNd();
1944 mImpl->setStrideNd(stride);
1954 return mImpl->getStrideNd();
1973 mImpl->setPaddingNd(padding);
1985 return mImpl->getPaddingNd();
1990 apiv::VPoolingLayer* mImpl;
2016 mImpl->setWindowSize(windowSize);
2026 return mImpl->getWindowSize();
2037 mImpl->setAlpha(alpha);
2047 return mImpl->getAlpha();
2058 mImpl->setBeta(beta);
2068 return mImpl->getBeta();
2089 return mImpl->getK();
2093 virtual ~
ILRNLayer() noexcept = default;
2094 apiv::VLRNLayer* mImpl;
2152 mImpl->setMode(mode);
2162 return mImpl->getMode();
2172 mImpl->setShift(shift);
2182 return mImpl->getShift();
2192 mImpl->setScale(scale);
2202 return mImpl->getScale();
2212 mImpl->setPower(power);
2222 return mImpl->getPower();
2237 return mImpl->getChannelAxis();
2258 mImpl->setChannelAxis(channelAxis);
2263 apiv::VScaleLayer* mImpl;
2312 mImpl->setAxes(axes);
2322 return mImpl->getAxes();
2327 apiv::VSoftMaxLayer* mImpl;
2357 mImpl->setAxis(axis);
2367 return mImpl->getAxis();
2372 apiv::VConcatenationLayer* mImpl;
2398 mImpl->setKernelSize(kernelSize);
2410 return mImpl->getKernelSize();
2422 mImpl->setNbOutputMaps(nbOutputMaps);
2432 return mImpl->getNbOutputMaps();
2449 mImpl->setStride(stride);
2461 return mImpl->getStride();
2481 mImpl->setPadding(padding);
2495 return mImpl->getPadding();
2515 mImpl->setNbGroups(nbGroups);
2525 return mImpl->getNbGroups();
2539 mImpl->setKernelWeights(weights);
2549 return mImpl->getKernelWeights();
2564 mImpl->setBiasWeights(weights);
2574 return mImpl->getBiasWeights();
2592 mImpl->setPrePadding(padding);
2602 return mImpl->getPrePadding();
2620 mImpl->setPostPadding(padding);
2630 return mImpl->getPostPadding();
2644 mImpl->setPaddingMode(paddingMode);
2656 return mImpl->getPaddingMode();
2671 mImpl->setKernelSizeNd(kernelSize);
2681 return mImpl->getKernelSizeNd();
2699 mImpl->setStrideNd(stride);
2709 return mImpl->getStrideNd();
2727 mImpl->setPaddingNd(padding);
2739 return mImpl->getPaddingNd();
2773 mImpl->setDilationNd(dilation);
2783 return mImpl->getDilationNd();
2788 apiv::VDeconvolutionLayer* mImpl;
2822 static constexpr int32_t kVALUE = 14;
2859 return mImpl->setOperation(op);
2871 return mImpl->getOperation();
2875 apiv::VElementWiseLayer* mImpl;
2990 mImpl->setGatherAxis(axis);
3001 return mImpl->getGatherAxis();
3020 mImpl->setNbElementWiseDims(elementWiseDims);
3030 return mImpl->getNbElementWiseDims();
3040 mImpl->setMode(mode);
3050 return mImpl->getMode();
3054 apiv::VGatherLayer* mImpl;
3240 return mImpl->getLayerCount();
3244 return mImpl->getHiddenSize();
3248 return mImpl->getMaxSeqLength();
3252 return mImpl->getDataLength();
3271 return mImpl->setSequenceLengths(seqLengths);
3283 return mImpl->getSequenceLengths();
3292 mImpl->setOperation(op);
3301 return mImpl->getOperation();
3310 mImpl->setInputMode(op);
3319 return mImpl->getInputMode();
3334 mImpl->setDirection(op);
3343 return mImpl->getDirection();
3402 mImpl->setWeightsForGate(layerIndex, gate, isW, weights);
3411 return mImpl->getWeightsForGate(layerIndex, gate, isW);
3436 mImpl->setBiasForGate(layerIndex, gate, isW, bias);
3445 return mImpl->getBiasForGate(layerIndex, gate, isW);
3462 mImpl->setHiddenState(hidden);
3471 return mImpl->getHiddenState();
3490 mImpl->setCellState(cell);
3499 return mImpl->getCellState();
3503 apiv::VRNNv2Layer* mImpl;
3526 return mImpl->getPlugin();
3530 apiv::VPluginV2Layer* mImpl;
3592 mImpl->setOperation(op);
3602 return mImpl->getOperation();
3606 apiv::VUnaryLayer* mImpl;
3661 mImpl->setOperation(op);
3671 return mImpl->getOperation();
3681 mImpl->setReduceAxes(reduceAxes);
3691 return mImpl->getReduceAxes();
3701 mImpl->setKeepDimensions(keepDimensions);
3711 return mImpl->getKeepDimensions();
3715 apiv::VReduceLayer* mImpl;
3743 mImpl->setPrePadding(padding);
3755 return mImpl->getPrePadding();
3769 mImpl->setPostPadding(padding);
3781 return mImpl->getPostPadding();
3795 mImpl->setPrePaddingNd(padding);
3807 return mImpl->getPrePaddingNd();
3821 mImpl->setPostPaddingNd(padding);
3833 return mImpl->getPostPaddingNd();
3837 apiv::VPaddingLayer* mImpl;
3878 mImpl->setFirstTranspose(permutation);
3890 return mImpl->getFirstTranspose();
3915 mImpl->setReshapeDimensions(dimensions);
3928 return mImpl->getReshapeDimensions();
3975 mImpl->setSecondTranspose(permutation);
3987 return mImpl->getSecondTranspose();
4003 return mImpl->setZeroIsPlaceholder(zeroIsPlaceholder);
4016 return mImpl->getZeroIsPlaceholder();
4020 apiv::VShuffleLayer* mImpl;
4091 mImpl->setStart(start);
4106 return mImpl->getStart();
4120 return mImpl->setSize(size);
4135 return mImpl->getSize();
4149 mImpl->setStride(stride);
4164 return mImpl->getStride();
4174 mImpl->setMode(mode);
4184 return mImpl->getMode();
4211 apiv::VSliceLayer* mImpl;
4230 apiv::VShapeLayer* mImpl;
4269 mImpl->setOperation(op);
4279 return mImpl->getOperation();
4301 return mImpl->getK();
4311 mImpl->setReduceAxes(reduceAxes);
4321 return mImpl->getReduceAxes();
4325 apiv::VTopKLayer* mImpl;
4400 mImpl->setOperation(index, op);
4410 return mImpl->getOperation(index);
4414 apiv::VMatrixMultiplyLayer* mImpl;
4435 apiv::VRaggedSoftMaxLayer* mImpl;
4454 apiv::VIdentityLayer* mImpl;
4479 mImpl->setWeights(weights);
4489 return mImpl->getWeights();
4501 mImpl->setDimensions(dimensions);
4513 return mImpl->getDimensions();
4517 apiv::VConstantLayer* mImpl;
4531 apiv::VParametricReLULayer* mImpl;
4552 static constexpr int32_t kVALUE = 2;
4602 static constexpr int32_t kVALUE = 3;
4628 static constexpr int32_t kVALUE = 2;
4660 static constexpr int32_t kVALUE = 4;
4708 return mImpl->setOutputDimensions(dimensions);
4718 return mImpl->getOutputDimensions();
4746 void setScales(
const float* scales, int32_t nbScales) noexcept
4748 mImpl->setScales(scales, nbScales);
4765 int32_t
getScales(int32_t size,
float* scales)
const noexcept
4767 return mImpl->getScales(size, scales);
4781 mImpl->setResizeMode(resizeMode);
4791 return mImpl->getResizeMode();
4808 mImpl->setAlignCorners(alignCorners);
4821 return mImpl->getAlignCorners();
4858 mImpl->setCoordinateTransformation(coordTransform);
4868 return mImpl->getCoordinateTransformation();
4885 mImpl->setSelectorForSinglePixel(selector);
4895 return mImpl->getSelectorForSinglePixel();
4911 mImpl->setNearestRounding(value);
4921 return mImpl->getNearestRounding();
4926 apiv::VResizeLayer* mImpl;
4972 return mBoundary->getLoop();
4977 apiv::VLoopBoundaryLayer* mBoundary;
4991 return mBoundary->getConditional();
4996 apiv::VConditionalBoundaryLayer* mBoundary;
5007 apiv::VConditionLayer* mImpl;
5020 apiv::VConditionalOutputLayer* mImpl;
5031 apiv::VConditionalInputLayer* mImpl;
5069 return mImpl->setCondition(condition);
5085 return mImpl->addOutput(trueSubgraphOutput, falseSubgraphOutput);
5097 return mImpl->addInput(input);
5110 mImpl->setName(name);
5120 return mImpl->getName();
5125 apiv::VIfConditional* mImpl;
5154 apiv::VRecurrenceLayer* mImpl;
5179 return mImpl->getLoopOutput();
5196 mImpl->setAxis(axis);
5202 return mImpl->getAxis();
5229 apiv::VLoopOutputLayer* mImpl;
5237 return mImpl->getTripLimit();
5242 apiv::VTripLimitLayer* mImpl;
5251 mImpl->setAxis(axis);
5257 return mImpl->getAxis();
5267 mImpl->setReverse(reverse);
5273 return mImpl->getReverse();
5278 apiv::VIteratorLayer* mImpl;
5297 return mImpl->addRecurrence(initialValue);
5318 return mImpl->addTripLimit(tensor, limit);
5331 return mImpl->addIterator(tensor, axis, reverse);
5343 return mImpl->addLoopOutput(tensor, outputKind, axis);
5356 mImpl->setName(name);
5366 return mImpl->getName();
5370 virtual ~
ILoop() noexcept = default;
5381 apiv::VSelectLayer* mImpl;
5411 mImpl->setMessage(message);
5421 return mImpl->getMessage();
5427 apiv::VAssertionLayer* mImpl;
5489 mImpl->setDimensions(dimensions);
5504 return mImpl->getDimensions();
5514 mImpl->setOperation(op);
5524 return mImpl->getOperation();
5542 mImpl->setAlpha(alpha);
5557 return mImpl->getAlpha();
5575 mImpl->setBeta(beta);
5590 return mImpl->getBeta();
5623 apiv::VFillLayer* mImpl;
5696 return mImpl->getAxis();
5707 mImpl->setAxis(axis);
5712 apiv::VQuantizeLayer* mImpl;
5783 return mImpl->getAxis();
5794 mImpl->setAxis(axis);
5799 apiv::VDequantizeLayer* mImpl;
5852 return mImpl->setEquation(equation);
5862 return mImpl->getEquation();
5867 apiv::VEinsumLayer* mImpl;
5954 mImpl->setMode(mode);
5964 return mImpl->getMode();
5974 mImpl->setAxis(axis);
5982 return mImpl->getAxis();
5986 apiv::VScatterLayer* mImpl;
6051 return mImpl->addInput(name, type, dimensions);
6065 mImpl->markOutput(tensor);
6089 return mImpl->addConvolution(input, nbOutputMaps, kernelSize, kernelWeights, biasWeights);
6110 return mImpl->addFullyConnected(input, nbOutputs, kernelWeights, biasWeights);
6129 return mImpl->addActivation(input, type);
6148 return mImpl->addPooling(input, type, windowSize);
6167 return mImpl->addLRN(input, window, alpha, beta, k);
6194 return mImpl->addScale(input, mode, shift, scale, power);
6207 return mImpl->addSoftMax(input);
6224 return mImpl->addConcatenation(inputs, nbInputs);
6248 return mImpl->addDeconvolution(input, nbOutputMaps, kernelSize, kernelWeights, biasWeights);
6273 return mImpl->addElementWise(input1, input2, op);
6293 return mImpl->addUnary(input, operation);
6310 return mImpl->addPadding(input, prePadding, postPadding);
6324 return mImpl->addShuffle(input);
6336 return mImpl->getNbLayers();
6350 return mImpl->getLayer(index);
6362 return mImpl->getNbInputs();
6378 return mImpl->getInput(index);
6392 return mImpl->getNbOutputs();
6408 return mImpl->getOutput(index);
6449 return mImpl->addReduce(input, operation, reduceAxes, keepDimensions);
6482 return mImpl->addTopK(input, op, k, reduceAxes);
6498 return mImpl->addGather(data, indices, axis);
6514 return mImpl->addGatherV2(data, indices, mode);
6532 return mImpl->addRaggedSoftMax(input, bounds);
6552 return mImpl->addMatrixMultiply(input0, op0, input1, op1);
6577 return mImpl->addConstant(dimensions, weights);
6644 ITensor& input, int32_t layerCount, int32_t hiddenSize, int32_t maxSeqLen,
RNNOperation op) noexcept
6646 return mImpl->addRNNv2(input, layerCount, hiddenSize, maxSeqLen, op);
6662 return mImpl->addIdentity(input);
6677 mImpl->removeTensor(tensor);
6689 mImpl->unmarkOutput(tensor);
6708 return mImpl->addPluginV2(inputs, nbInputs, plugin);
6727 return mImpl->addSlice(input, start, size, stride);
6749 mImpl->setName(name);
6763 return mImpl->getName();
6781 return mImpl->addShape(input);
6800 return mImpl->hasImplicitBatchDimension();
6818 return mImpl->markOutputForShapes(tensor);
6830 return mImpl->unmarkOutputForShapes(tensor);
6848 return mImpl->addParametricReLU(input, slope);
6871 return mImpl->addConvolutionNd(input, nbOutputMaps, kernelSize, kernelWeights, biasWeights);
6890 return mImpl->addPoolingNd(input, type, windowSize);
6913 return mImpl->addDeconvolutionNd(input, nbOutputMaps, kernelSize, kernelWeights, biasWeights);
6949 return mImpl->addScaleNd(input, mode, shift, scale, power, channelAxis);
6964 return mImpl->addResize(input);
6981 return mImpl->hasExplicitPrecision();
6997 return mImpl->addLoop();
7035 return mImpl->addSelect(condition, thenInput, elseInput);
7052 return mImpl->addAssertion(condition, message);
7070 return mImpl->addFill(dimensions, op);
7087 return mImpl->addPaddingNd(input, prePadding, postPadding);
7107 return mImpl->setWeightsName(weights, name);
7126 mImpl->setErrorRecorder(recorder);
7141 return mImpl->getErrorRecorder();
7160 return mImpl->addDequantize(input, scale);
7180 return mImpl->addScatter(data, indices, updates, mode);
7199 return mImpl->addQuantize(input, scale);
7214 return mImpl->addIfConditional();
7228 return mImpl->addEinsum(inputs, nbInputs, equation);
7232 apiv::VNetworkDefinition* mImpl;
7242 kLEGACY_CALIBRATION = 0,
7243 kENTROPY_CALIBRATION = 1,
7244 kENTROPY_CALIBRATION_2 = 2,
7245 kMINMAX_CALIBRATION = 3,
7274 virtual int32_t getBatchSize()
const noexcept = 0;
7289 virtual bool getBatch(
void* bindings[],
const char* names[], int32_t nbBindings) noexcept = 0;
7305 virtual const void* readCalibrationCache(std::size_t& length) noexcept = 0;
7315 virtual void writeCalibrationCache(
const void* ptr, std::size_t length) noexcept = 0;
7339 return CalibrationAlgoType::kENTROPY_CALIBRATION;
7357 return CalibrationAlgoType::kENTROPY_CALIBRATION_2;
7374 return CalibrationAlgoType::kMINMAX_CALIBRATION;
7392 return CalibrationAlgoType::kLEGACY_CALIBRATION;
7401 virtual double getQuantile() const noexcept = 0;
7409 virtual
double getRegressionCutoff() const noexcept = 0;
7423 virtual const
void* readHistogramCache(std::
size_t& length) noexcept = 0;
7433 virtual
void writeHistogramCache(const
void* ptr, std::
size_t length) noexcept = 0;
7456 return mImpl->getTensorFormat();
7464 return mImpl->getDataType();
7472 return mImpl->getStrides();
7477 apiv::VAlgorithmIOInfo* mImpl;
7499 return mImpl->getImplementation();
7507 return mImpl->getTactic();
7512 apiv::VAlgorithmVariant* mImpl;
7532 return mImpl->getName();
7543 return mImpl->getDimensions(index, select);
7551 return mImpl->getNbInputs();
7559 return mImpl->getNbOutputs();
7564 apiv::VAlgorithmContext* mImpl;
7591 return mImpl->getAlgorithmIOInfo(index);
7599 return mImpl->getAlgorithmVariant();
7607 return mImpl->getTimingMSec();
7615 return mImpl->getWorkspaceSize();
7628 return mImpl->getAlgorithmIOInfoByIndex(index);
7633 apiv::VAlgorithm* mImpl;
7661 int32_t nbChoices, int32_t* selection) noexcept
7674 int32_t nbAlgorithms) noexcept
7784 return mImpl->serialize();
7808 return mImpl->combine(inputCache, ignoreMismatch);
7818 return mImpl->reset();
7822 apiv::VTimingCache* mImpl;
7847 mImpl->setMinTimingIterations(minTiming);
7859 return mImpl->getMinTimingIterations();
7872 mImpl->setAvgTimingIterations(avgTiming);
7884 return mImpl->getAvgTimingIterations();
7897 mImpl->setEngineCapability(capability);
7909 return mImpl->getEngineCapability();
7919 mImpl->setInt8Calibrator(calibrator);
7927 return mImpl->getInt8Calibrator();
7939 mImpl->setMaxWorkspaceSize(workspaceSize);
7953 return mImpl->getMaxWorkspaceSize();
7970 mImpl->setFlags(builderFlags);
7982 return mImpl->getFlags();
7994 mImpl->clearFlag(builderFlag);
8006 mImpl->setFlag(builderFlag);
8018 return mImpl->getFlag(builderFlag);
8033 mImpl->setDeviceType(layer, deviceType);
8042 return mImpl->getDeviceType(layer);
8052 return mImpl->isDeviceTypeSet(layer);
8062 mImpl->resetDeviceType(layer);
8071 return mImpl->canRunOnDLA(layer);
8086 mImpl->setDLACore(dlaCore);
8097 return mImpl->getDLACore();
8107 mImpl->setDefaultDeviceType(deviceType);
8117 return mImpl->getDefaultDeviceType();
8153 return mImpl->setProfileStream(stream);
8165 return mImpl->getProfileStream();
8181 return mImpl->addOptimizationProfile(profile);
8194 return mImpl->getNbOptimizationProfiles();
8206 mImpl->setProfilingVerbosity(verbosity);
8219 return mImpl->getProfilingVerbosity();
8228 mImpl->setAlgorithmSelector(selector);
8236 return mImpl->getAlgorithmSelector();
8251 return mImpl->setCalibrationProfile(profile);
8261 return mImpl->getCalibrationProfile();
8278 mImpl->setQuantizationFlags(flags);
8290 return mImpl->getQuantizationFlags();
8302 mImpl->clearQuantizationFlag(flag);
8314 mImpl->setQuantizationFlag(flag);
8326 return mImpl->getQuantizationFlag(flag);
8351 return mImpl->setTacticSources(tacticSources);
8366 return mImpl->getTacticSources();
8385 return mImpl->createTimingCache(blob, size);
8408 return mImpl->setTimingCache(cache, ignoreMismatch);
8418 return mImpl->getTimingCache();
8422 apiv::VBuilderConfig* mImpl;
8482 virtual ~
IBuilder() noexcept =
default;
8494 mImpl->setMaxBatchSize(batchSize);
8507 return mImpl->getMaxBatchSize();
8515 return mImpl->platformHasFastFp16();
8523 return mImpl->platformHasFastInt8();
8547 return mImpl->getMaxDLABatchSize();
8555 return mImpl->getNbDLACores();
8571 mImpl->setGpuAllocator(allocator);
8581 return mImpl->createBuilderConfig();
8597 return mImpl->buildEngineWithConfig(network, config);
8613 return mImpl->createNetworkV2(flags);
8627 return mImpl->createOptimizationProfile();
8646 mImpl->setErrorRecorder(recorder);
8661 return mImpl->getErrorRecorder();
8677 return mImpl->platformHasTf32();
8696 return mImpl->buildSerializedNetwork(network, config);
8718 return mImpl->isNetworkSupported(network, config);
8728 return mImpl->getLogger();
8732 apiv::VBuilder* mImpl;
8741 extern "C" TENSORRTAPI
void* createInferBuilder_INTERNAL(
void* logger, int32_t version) noexcept;
8757 return static_cast<IBuilder*>(createInferBuilder_INTERNAL(&logger, NV_TENSORRT_VERSION));
8763 #endif // NV_INFER_H
Dims getStrideNd() const noexcept
Get the multi-dimension stride of the deconvolution.
Definition: NvInfer.h:2707
void removeTensor(ITensor &tensor) noexcept
remove a tensor from the network definition.
Definition: NvInfer.h:6675
TRT_DEPRECATED void setKernelSize(DimsHW kernelSize) noexcept
Set the HW kernel size of the convolution.
Definition: NvInfer.h:1005
void setDimensions(Dims dimensions) noexcept
Set the output tensor's dimensions.
Definition: NvInfer.h:5487
LoopOutput
Enum that describes kinds of loop outputs.
Definition: NvInfer.h:4930
TRT_DEPRECATED DimsHW getPadding() const noexcept
Get the padding for pooling.
Definition: NvInfer.h:1768
float getAlpha() const noexcept
Get the alpha parameter.
Definition: NvInfer.h:1612
Dims getStrideNd() const noexcept
Get the multi-dimension stride of the convolution.
Definition: NvInfer.h:1332
IErrorRecorder * getErrorRecorder() const noexcept
get the ErrorRecorder assigned to this interface.
Definition: NvInfer.h:7139
TRT_DEPRECATED void destroy() noexcept
Destroy this INetworkDefinition object.
Definition: NvInfer.h:6418
void setStride(Dims stride) noexcept
Set the stride for computing the output slice data.
Definition: NvInfer.h:4147
ITensor * addInput(const char *name, DataType type, Dims dimensions) noexcept
Add an input tensor to the network.
Definition: NvInfer.h:6049
Dims getPostPadding() const noexcept
Get the padding.
Definition: NvInfer.h:2628
Hard sigmoid activation: max(0, min(1, alpha*x+beta))
void setDilationNd(Dims dilation) noexcept
Set the multi-dimension dilation of the convolution.
Definition: NvInfer.h:1376
void setReduceAxes(uint32_t reduceAxes) noexcept
Set which axes to reduce for the layer.
Definition: NvInfer.h:4309
TRT_DEPRECATED DimsHW getPadding() const noexcept
Get the padding of the deconvolution.
Definition: NvInfer.h:2493
void setWindowSize(int32_t windowSize) noexcept
Set the LRN window size.
Definition: NvInfer.h:2014
Use explicit padding, rounding output size up.
TRT_DEPRECATED void setAlignCorners(bool alignCorners) noexcept
Set whether to align corners while resizing.
Definition: NvInfer.h:4806
void setAlgorithmSelector(IAlgorithmSelector *selector) noexcept
Set Algorithm Selector.
Definition: NvInfer.h:8226
void setPrePaddingNd(Dims padding) noexcept
Set the padding that is applied at the start of the tensor.
Definition: NvInfer.h:3793
void setFirstTranspose(Permutation permutation) noexcept
Set the permutation applied by the first transpose operation.
Definition: NvInfer.h:3876
SliceMode
Controls how ISliceLayer handles out of bounds coordinates.
Definition: NvInfer.h:4029
Use SAME padding, with prePadding >= postPadding.
void reset() noexcept
Resets the builder configuration to defaults.
Definition: NvInfer.h:8125
TRT_DEPRECATED void setStride(DimsHW stride) noexcept
Get the stride of the convolution.
Definition: NvInfer.h:1055
Parametric softplus activation: alpha*log(exp(beta*x)+1)
UnaryOperation getOperation() const noexcept
Get the unary operation for the layer.
Definition: NvInfer.h:3600
void setFlag(BuilderFlag builderFlag) noexcept
Set a single build mode flag.
Definition: NvInfer.h:8004
void setPaddingNd(Dims padding) noexcept
Set the multi-dimension padding of the deconvolution.
Definition: NvInfer.h:2725
void setOperation(RNNOperation op) noexcept
Set the operation of the RNN layer.
Definition: NvInfer.h:3290
A fully connected layer in a network definition. This layer expects an input tensor of three or more ...
Definition: NvInfer.h:1448
Definition: NvInfer.h:3841
ISoftMaxLayer * addSoftMax(ITensor &input) noexcept
Add a SoftMax layer to the network.
Definition: NvInfer.h:6205
void setPostPaddingNd(Dims padding) noexcept
Set the padding that is applied at the end of the tensor.
Definition: NvInfer.h:3819
No operation is performed on the first recurrent layer.
TRT_DEPRECATED DimsHW getKernelSize() const noexcept
Get the HW kernel size of the convolution.
Definition: NvInfer.h:1017
Layer that represents a constant value.
Definition: NvInfer.h:4465
Use CAFFE padding, rounding output size down, uses prePadding value.
TRT_DEPRECATED IDeconvolutionLayer * addDeconvolution(ITensor &input, int32_t nbOutputMaps, DimsHW kernelSize, Weights kernelWeights, Weights biasWeights) noexcept
Add a deconvolution layer to the network.
Definition: NvInfer.h:6245
Layer that represents a parametric ReLU operation.
Definition: NvInfer.h:4528
IPluginV2Layer * addPluginV2(ITensor *const *inputs, int32_t nbInputs, IPluginV2 &plugin) noexcept
Add a plugin layer to the network using the IPluginV2 interface.
Definition: NvInfer.h:6706
const IOptimizationProfile * getCalibrationProfile() noexcept
Get the current calibration profile.
Definition: NvInfer.h:8259
IIfConditional * getConditional() const noexcept
Return pointer to the IIfConditional associated with this boundary layer.
Definition: NvInfer.h:4989
constexpr int32_t EnumMax< RNNDirection >() noexcept
Maximum number of elements in RNNDirection enum.
Definition: NvInfer.h:3167
TRT_DEPRECATED IPaddingLayer * addPadding(ITensor &input, DimsHW prePadding, DimsHW postPadding) noexcept
Add a padding layer to the network.
Definition: NvInfer.h:6308
Minimum of the two elements.
IConvolutionLayer * addConvolutionNd(ITensor &input, int32_t nbOutputMaps, Dims kernelSize, Weights kernelWeights, Weights biasWeights) noexcept
Add a multi-dimension convolution layer to the network.
Definition: NvInfer.h:6868
IErrorRecorder * getErrorRecorder() const noexcept
get the ErrorRecorder assigned to this interface.
Definition: NvInfer.h:8659
uint32_t getAxes() const noexcept
Get the axis along which softmax occurs.
Definition: NvInfer.h:2320
A network definition for input to the builder.
Definition: NvInfer.h:6009
bool getQuantizationFlag(QuantizationFlag flag) const noexcept
Returns true if the quantization flag is set.
Definition: NvInfer.h:8324
bool getFlag(BuilderFlag builderFlag) const noexcept
Returns true if the build mode flag is set.
Definition: NvInfer.h:8016
void setAxis(int32_t axis) noexcept
Set the quantization axis.
Definition: NvInfer.h:5705
void setNbElementWiseDims(int32_t elementWiseDims) noexcept
Set the number of leading dimensions of indices tensor to be handled elementwise. The gathering of in...
Definition: NvInfer.h:3018
Per-channel coefficients.
void setPostPadding(Dims padding) noexcept
Set the multi-dimension post-padding for pooling.
Definition: NvInfer.h:1870
void setDimensions(Dims dimensions) noexcept
Set the dimensions of a tensor.
Definition: NvInfer.h:239
An array of weights used as a layer parameter.
Definition: NvInferRuntime.h:155
provides a unique 128-bit identifier, which along with the input and output information denotes the v...
Definition: NvInfer.h:7491
int32_t getGatherAxis() const noexcept
Get the axis to gather on.
Definition: NvInfer.h:2999
bool isShapeTensor() const noexcept
Whether the tensor is a shape tensor.
Definition: NvInfer.h:477
bool combine(const ITimingCache &inputCache, bool ignoreMismatch) noexcept
Combine input timing cache into local instance.
Definition: NvInfer.h:7806
int32_t getNbGroups() const noexcept
Get the number of groups of the convolution.
Definition: NvInfer.h:1127
Layer that represents a padding operation.
Definition: NvInfer.h:3729
bool canRunOnDLA(const ILayer *layer) const noexcept
Checks if a layer can run on DLA.
Definition: NvInfer.h:8069
IFullyConnectedLayer * addFullyConnected(ITensor &input, int32_t nbOutputs, Weights kernelWeights, Weights biasWeights) noexcept
Add a fully connected layer to the network.
Definition: NvInfer.h:6107
A convolution layer in a network definition.
Definition: NvInfer.h:993
bool setDynamicRange(float min, float max) noexcept
Set dynamic range for the tensor.
Definition: NvInfer.h:294
DataType getOutputType(int32_t index) const noexcept
get the output type of this layer
Definition: NvInfer.h:704
TRT_DEPRECATED void setStride(DimsHW stride) noexcept
Set the stride for pooling.
Definition: NvInfer.h:1726
void setActivationType(ActivationType type) noexcept
Set the type of activation to be performed.
Definition: NvInfer.h:1564
void setDirection(RNNDirection op) noexcept
Set the direction of the RNN layer.
Definition: NvInfer.h:3332
Class to handle library allocated memory that is accessible to the user.
Definition: NvInferRuntime.h:173
Dims getOutputDimensions() const noexcept
Get the output dimensions.
Definition: NvInfer.h:4716
A layer that represents the identity function.
Definition: NvInfer.h:4451
void setPostPadding(Dims padding) noexcept
Set the multi-dimension post-padding of the convolution.
Definition: NvInfer.h:1248
RNNOperation getOperation() const noexcept
Get the operation of the RNN layer.
Definition: NvInfer.h:3299
DeviceType
The device that this layer/network will execute on.
Definition: NvInferRuntime.h:631
int32_t getNbOutputs() const noexcept
Get the number of outputs in the network.
Definition: NvInfer.h:6390
Network iterates from first to last and vice versa and outputs concatenated.
ProfilingVerbosity
List of verbosity levels of layer information exposed in NVTX annotations and in IEngineInspector.
Definition: NvInferRuntime.h:1285
const IAlgorithmIOInfo * getAlgorithmIOInfoByIndex(int32_t index) const noexcept
Returns the format of an Algorithm input or output. Algorithm inputs are incrementally numbered first...
Definition: NvInfer.h:7626
void setAxes(uint32_t axes) noexcept
Set the axis along which softmax is computed. Currently, only one axis can be set.
Definition: NvInfer.h:2310
Selu activation: x>0 ? beta * x : beta * (alpha*exp(x) - alpha)
CalibrationAlgoType getAlgorithm() noexcept override
Definition: NvInfer.h:7337
Interface implemented by application for selecting and reporting algorithms of a layer provided by th...
Definition: NvInfer.h:7644
RNNInputMode
Enumerates the RNN input modes that may occur with an RNN layer.
Definition: NvInfer.h:3187
void setBiasWeights(Weights weights) noexcept
Set the bias weights.
Definition: NvInfer.h:1500
int32_t getNbOptimizationProfiles() const noexcept
Get number of optimization profiles.
Definition: NvInfer.h:8192
void clearFlag(BuilderFlag builderFlag) noexcept
clear a single build mode flag.
Definition: NvInfer.h:7992
Declaration of EnumMaxImpl struct to store maximum number of elements in an enumeration type.
Definition: NvInferRuntimeCommon.h:136
int32_t getAvgTimingIterations() const noexcept
Query the number of averaging iterations.
Definition: NvInfer.h:7882
IDequantizeLayer * addDequantize(ITensor &input, ITensor &scale) noexcept
Add a dequantization layer to the network.
Definition: NvInfer.h:7158
void setDilationNd(Dims dilation) noexcept
Set the multi-dimension dilation of the deconvolution.
Definition: NvInfer.h:2771
Dims getPostPaddingNd() const noexcept
Get the padding that is applied at the end of the tensor.
Definition: NvInfer.h:3831
TopKOperation getOperation() const noexcept
Get the operation for the layer.
Definition: NvInfer.h:4277
IIfConditionalInputLayer * addInput(ITensor &input) noexcept
Add an If-conditional input.
Definition: NvInfer.h:5095
int32_t getNbGroups() const noexcept
Get the number of groups for a deconvolution.
Definition: NvInfer.h:2523
bool hasImplicitBatchDimension() const noexcept
Query whether the network was created with an implicit batch dimension.
Definition: NvInfer.h:6798
int32_t getScales(int32_t size, float *scales) const noexcept
Copies resize scales to scales[0, ..., nbScales-1], where nbScales is the number of scales that were ...
Definition: NvInfer.h:4765
Definition: NvInfer.h:5232
Clip activation: max(alpha, min(beta, x))
Permutation getFirstTranspose() const noexcept
Get the permutation applied by the first transpose operation.
Definition: NvInfer.h:3888
int32_t addOptimizationProfile(const IOptimizationProfile *profile) noexcept
Add an optimization profile.
Definition: NvInfer.h:8179
IElementWiseLayer * addElementWise(ITensor &input1, ITensor &input2, ElementWiseOperation op) noexcept
Add an elementwise layer to the network.
Definition: NvInfer.h:6271
GatherMode getMode() const noexcept
Get the gather mode.
Definition: NvInfer.h:3048
Plugin class for user-implemented layers.
Definition: NvInferRuntimeCommon.h:409
Carries information about input or output of the algorithm. IAlgorithmIOInfo for all the input and ou...
Definition: NvInfer.h:7448
float getTimingMSec() const noexcept
The time in milliseconds to execute the algorithm.
Definition: NvInfer.h:7605
void setK(float k) noexcept
Set the LRN K value.
Definition: NvInfer.h:2077
Definition: NvInferRuntimeCommon.h:189
Definition: NvInfer.h:7366
TensorFormat getTensorFormat() const noexcept
Return TensorFormat of the input/output of algorithm.
Definition: NvInfer.h:7454
void markOutput(ITensor &tensor) noexcept
Mark a tensor as a network output.
Definition: NvInfer.h:6063
Scaled tanh activation: alpha*tanh(beta*x)
constexpr int32_t EnumMax< FillOperation >() noexcept
Maximum number of elements in FillOperation enum.
Definition: NvInfer.h:5445
int32_t getNbInputs() const noexcept
Return number of inputs of the algorithm.
Definition: NvInfer.h:7549
void setType(DataType type) noexcept
Set the data type of a tensor.
Definition: NvInfer.h:267
void setPoolingType(PoolingType type) noexcept
Set the type of activation to be performed.
Definition: NvInfer.h:1674
Softsign activation: x / (1+|x|)
Out of bounds indices are clamped to bounds.
Enable layers marked to execute on GPU if layer cannot execute on DLA.
void resetDynamicRange() noexcept
Undo effect of setDynamicRange.
Definition: NvInfer.h:391
TRT_DEPRECATED void destroy() noexcept
Destroy this object.
Definition: NvInfer.h:8533
IRaggedSoftMaxLayer * addRaggedSoftMax(ITensor &input, ITensor &bounds) noexcept
Add a RaggedSoftMax layer to the network.
Definition: NvInfer.h:6530
void setLocation(TensorLocation location) noexcept
Set the storage location of a tensor.
Definition: NvInfer.h:373
Like kNONE, but transpose the matrix dimensions.
Enable building a refittable engine.
ScatterMode getMode() const noexcept
Get the scatter mode.
Definition: NvInfer.h:5962
nvinfer1::IHostMemory * buildSerializedNetwork(INetworkDefinition &network, IBuilderConfig &config) noexcept
Builds and serializes a network for the given INetworkDefinition and IBuilderConfig.
Definition: NvInfer.h:8694
ISelectLayer * addSelect(ITensor &condition, ITensor &thenInput, ITensor &elseInput) noexcept
Add a select layer to the network.
Definition: NvInfer.h:7033
void setPrePadding(Dims padding) noexcept
Set the multi-dimension pre-padding of the deconvolution.
Definition: NvInfer.h:2590
An RNN layer in a network definition, version 2.
Definition: NvInfer.h:3235
float getAlpha() const noexcept
Get the LRN alpha value.
Definition: NvInfer.h:2045
DataType getDataType() const noexcept
Return DataType of the input/output of algorithm.
Definition: NvInfer.h:7462
void setBroadcastAcrossBatch(bool broadcastAcrossBatch) noexcept
Set whether to enable broadcast of tensor across the batch.
Definition: NvInfer.h:332
void setPaddingMode(PaddingMode paddingMode) noexcept
Set the padding mode.
Definition: NvInfer.h:1893
Logical AND of two elements.
int32_t getNbOutputChannels() const noexcept
Get the number of output channels K from the fully connected layer.
Definition: NvInfer.h:1468
Enable Int8 layer selection, with FP32 fallback with FP16 fallback if kFP16 also specified.
Dims getPostPadding() const noexcept
Get the post-padding.
Definition: NvInfer.h:1258
TRT_DEPRECATED bool getAlignCorners() const noexcept
True if align corners has been set.
Definition: NvInfer.h:4819
Similar to ONNX GatherND.
TRT_DEPRECATED DimsHW getWindowSize() const noexcept
Get the window size for pooling.
Definition: NvInfer.h:1710
IRecurrenceLayer * addRecurrence(ITensor &initialValue) noexcept
Create a recurrence layer for this loop with initialValue as its first input.
Definition: NvInfer.h:5295
Dims getDimensions() const noexcept
Get the output tensor's dimensions.
Definition: NvInfer.h:5502
Descriptor for two-dimensional spatial data.
Definition: NvInferLegacyDims.h:100
cudaStream_t getProfileStream() const noexcept
Get the cuda stream that is used to profile this network.
Definition: NvInfer.h:8163
Dims getStrideNd() const noexcept
Get the multi-dimension stride for pooling.
Definition: NvInfer.h:1952
Dims getReshapeDimensions() const noexcept
Get the reshaped dimensions.
Definition: NvInfer.h:3926
bool platformHasFastInt8() const noexcept
Determine whether the platform has fast native int8.
Definition: NvInfer.h:8521
IConditionLayer * setCondition(ITensor &condition) noexcept
Set the condition tensor for this If-Conditional construct.
Definition: NvInfer.h:5067
ResizeMode
Enumerates various modes of resize in the resize layer. Resize mode set using setResizeMode().
Definition: NvInfer.h:4540
TensorFormat
Format of the input/output tensors.
Definition: NvInferRuntimeCommon.h:220
void setWeightsForGate(int32_t layerIndex, RNNGateType gate, bool isW, Weights weights) noexcept
Set the weight parameters for an individual gate in the RNN.
Definition: NvInfer.h:3400
Definition: NvInfer.h:5174
constexpr int32_t EnumMax< MatrixOperation >() noexcept
Maximum number of elements in MatrixOperation enum.
Definition: NvInfer.h:4359
IUnaryLayer * addUnary(ITensor &input, UnaryOperation operation) noexcept
Add a unary layer to the network.
Definition: NvInfer.h:6291
void setBiasWeights(Weights weights) noexcept
Set the bias weights for the convolution.
Definition: NvInfer.h:1166
IDeconvolutionLayer * addDeconvolutionNd(ITensor &input, int32_t nbOutputMaps, Dims kernelSize, Weights kernelWeights, Weights biasWeights) noexcept
Add a multi-dimension deconvolution layer to the network.
Definition: NvInfer.h:6910
nvinfer1::INetworkDefinition * createNetworkV2(NetworkDefinitionCreationFlags flags) noexcept
Create a network definition object.
Definition: NvInfer.h:8611
Dims getStrides() const noexcept
Return strides of the input/output tensor of algorithm.
Definition: NvInfer.h:7470
void setPaddingNd(Dims padding) noexcept
Set the multi-dimension padding for pooling.
Definition: NvInfer.h:1971
int32_t getNbOutputs() const noexcept
Get the number of outputs of a layer.
Definition: NvInfer.h:577
PoolingType getPoolingType() const noexcept
Get the type of activation to be performed.
Definition: NvInfer.h:1684
void setNbOutputMaps(int32_t nbOutputMaps) noexcept
Set the number of output feature maps for the deconvolution.
Definition: NvInfer.h:2420
bool markOutputForShapes(ITensor &tensor) noexcept
Enable tensor's value to be computed by IExecutionContext::getShapeBinding.
Definition: NvInfer.h:6816
Output value is concatenation of values of tensor for each iteration, in reverse order.
bool getKeepDimensions() const noexcept
Get the boolean that specifies whether or not to keep the reduced dimensions for the layer.
Definition: NvInfer.h:3709
Application-implemented logging interface for the builder, refitter and runtime.
Definition: NvInferRuntimeCommon.h:1256
Describes the context and requirements, that could be fulfilled by one or more instances of IAlgorith...
Definition: NvInfer.h:7523
Dims getStart() const noexcept
Get the start offset for the slice layer.
Definition: NvInfer.h:4104
bool outputTypeIsSet(int32_t index) const noexcept
whether the output type has been set for this layer
Definition: NvInfer.h:717
Dims getPrePadding() const noexcept
Get the pre-padding.
Definition: NvInfer.h:1231
TRT_DEPRECATED DimsHW getPadding() const noexcept
Get the padding of the convolution. If the padding is asymmetric, the pre-padding is returned.
Definition: NvInfer.h:1097
constexpr int32_t EnumMax< QuantizationFlag >() noexcept
Maximum number of quantization flags in QuantizationFlag enum.
Definition: NvInfer.h:7705
void setChannelAxis(int32_t channelAxis) noexcept
Set the channel axis.
Definition: NvInfer.h:2256
void setReverse(bool reverse) noexcept
Definition: NvInfer.h:5265
IShuffleLayer * addShuffle(ITensor &input) noexcept
Add a shuffle layer to the network.
Definition: NvInfer.h:6322
void unmarkOutput(ITensor &tensor) noexcept
unmark a tensor as a network output.
Definition: NvInfer.h:6687
A scatter layer in a network definition. Supports several kinds of scattering.
Definition: NvInfer.h:5944
An engine for executing inference on a built network, with functionally unsafe features.
Definition: NvInferRuntime.h:1308
Enable FP16 layer selection, with FP32 fallback.
ActivationType
Enumerates the types of activation to perform in an activation layer.
Definition: NvInfer.h:154
void setPrecision(DataType dataType) noexcept
Set the computational precision of this layer.
Definition: NvInfer.h:625
Dims getWindowSizeNd() const noexcept
Get the multi-dimension window size for pooling.
Definition: NvInfer.h:1927
Permutation getSecondTranspose() const noexcept
Get the permutation applied by the second transpose operation.
Definition: NvInfer.h:3985
A RaggedSoftmax layer in a network definition.
Definition: NvInfer.h:4432
void setKeepDimensions(bool keepDimensions) noexcept
Set the boolean that specifies whether or not to keep the reduced dimensions for the layer.
Definition: NvInfer.h:3699
void setScale(Weights scale) noexcept
Set the scale value.
Definition: NvInfer.h:2190
Dims getDilationNd() const noexcept
Get the multi-dimension dilation of the convolution.
Definition: NvInfer.h:1386
void setReshapeDimensions(Dims dimensions) noexcept
Set the reshaped dimensions.
Definition: NvInfer.h:3913
void clearQuantizationFlag(QuantizationFlag flag) noexcept
clear a quantization flag.
Definition: NvInfer.h:8300
float getK() const noexcept
Get the LRN K value.
Definition: NvInfer.h:2087
Use SAME padding, with prePadding <= postPadding.
int32_t getAxis() const noexcept
Get axis being concatenated over.
Definition: NvInfer.h:5200
const char * getEquation() const noexcept
Return the equation.
Definition: NvInfer.h:5860
constexpr int32_t EnumMax< RNNInputMode >() noexcept
Maximum number of elements in RNNInputMode enum.
Definition: NvInfer.h:3195
void setKernelWeights(Weights weights) noexcept
Set the kernel weights, given as a KxC matrix in row-major order.
Definition: NvInfer.h:1478
RNNDirection getDirection() const noexcept
Get the direction of the RNN layer.
Definition: NvInfer.h:3341
Dims getDimensions(int32_t index, OptProfileSelector select) const noexcept
Get the minimum / optimum / maximum dimensions for input or output tensor.
Definition: NvInfer.h:7541
void setKernelWeights(Weights weights) noexcept
Set the kernel weights for the deconvolution.
Definition: NvInfer.h:2537
float getDynamicRangeMin() const noexcept
Get minimum of dynamic range.
Definition: NvInfer.h:401
ElementWiseOperation
Enumerates the binary operations that may be performed by an ElementWise layer.
Definition: NvInfer.h:2798
Generate a tensor with random values drawn from a uniform distribution.
ScaleMode getMode() const noexcept
Get the scale mode.
Definition: NvInfer.h:2160
IQuantizeLayer * addQuantize(ITensor &input, ITensor &scale) noexcept
Add a quantization layer to the network.
Definition: NvInfer.h:7197
Divide the first element by the second.
int32_t getChannelAxis() const noexcept
Get the channel axis.
Definition: NvInfer.h:2235
IInt8Calibrator * getInt8Calibrator() const noexcept
Get Int8 Calibration interface.
Definition: NvInfer.h:7925
Weights getKernelWeights() const noexcept
Get the kernel weights for the deconvolution.
Definition: NvInfer.h:2547
Slices an input tensor into an output tensor based on the offset and strides.
Definition: NvInfer.h:4077
TRT_DEPRECATED DimsHW getDilation() const noexcept
Get the dilation for a convolution.
Definition: NvInfer.h:1204
void setOperation(int32_t index, MatrixOperation op) noexcept
Set the operation for an input tensor.
Definition: NvInfer.h:4398
void setAllowedFormats(TensorFormats formats) noexcept
Set allowed formats for this tensor. By default all formats are allowed. Shape tensors (for which isS...
Definition: NvInfer.h:430
TRT_DEPRECATED void setWindowSize(DimsHW windowSize) noexcept
Set the window size for pooling.
Definition: NvInfer.h:1698
ScaleMode
Controls how shift, scale and power are applied in a Scale layer.
Definition: NvInfer.h:2102
TacticSources getTacticSources() const noexcept
Get tactic sources.
Definition: NvInfer.h:8364
Use explicit padding, rounding output size down.
std::size_t getWorkspaceSize() const noexcept
The size of the GPU temporary memory in bytes which the algorithm uses at execution time.
Definition: NvInfer.h:7613
Logical OR of two elements.
IConstantLayer * addConstant(Dims dimensions, Weights weights) noexcept
Add a constant layer to the network.
Definition: NvInfer.h:6575
ReduceOperation
Enumerates the reduce operations that may be performed by a Reduce layer.
Definition: NvInfer.h:3628
Rectified linear activation.
constexpr int32_t EnumMax< SliceMode >() noexcept
Maximum number of elements in SliceMode enum.
Definition: NvInfer.h:4042
Definition: NvInfer.h:5245
void setProfilingVerbosity(ProfilingVerbosity verbosity) noexcept
Set verbosity level of layer information exposed in NVTX annotations and IEngineInspector.
Definition: NvInfer.h:8204
void setZeroIsPlaceholder(bool zeroIsPlaceholder) noexcept
Set meaning of 0 in reshape dimensions.
Definition: NvInfer.h:4001
void setOutputDimensions(Dims dimensions) noexcept
Set the output dimensions.
Definition: NvInfer.h:4706
bool unmarkOutputForShapes(ITensor &tensor) noexcept
Undo markOutputForShapes.
Definition: NvInfer.h:6828
int32_t getNbLayers() const noexcept
Get the number of layers in the network.
Definition: NvInfer.h:6334
BuilderFlag
List of valid modes that the builder can enable when creating an engine from a network definition.
Definition: NvInfer.h:7725
void setAlpha(float alpha) noexcept
Set the alpha parameter (must be finite).
Definition: NvInfer.h:1589
void setBeta(float beta) noexcept
Set the beta parameter (must be finite).
Definition: NvInfer.h:1603
void setDefaultDeviceType(DeviceType deviceType) noexcept
Sets the default DeviceType to be used by the builder. It ensures that all the layers that can run on...
Definition: NvInfer.h:8105
constexpr int32_t EnumMax< TopKOperation >() noexcept
Maximum number of elements in TopKOperation enum.
Definition: NvInfer.h:4247
The first element to the power of the second element.
PaddingMode getPaddingMode() const noexcept
Get the padding mode.
Definition: NvInfer.h:2654
bool getReverse() const noexcept
True if and only if reversing input.
Definition: NvInfer.h:5271
void setAxis(int32_t axis) noexcept
Set where to insert the contenation axis. Ignored if getLoopOutput() is kLAST_VALUE.
Definition: NvInfer.h:5194
void setBlendFactor(float blendFactor) noexcept
Set the blending factor for the max_average_blend mode: max_average_blendPool = (1-blendFactor)*maxPo...
Definition: NvInfer.h:1783
Weights getBiasWeights() const noexcept
Get the bias weights for the deconvolution.
Definition: NvInfer.h:2572
constexpr int32_t EnumMax< ScaleMode >() noexcept
Maximum number of elements in ScaleMode enum.
Definition: NvInfer.h:2111
Sign, If input > 0, output 1; if input < 0, output -1; if input == 0, output 0.
ElementWiseOperation getOperation() const noexcept
Get the binary operation for the layer.
Definition: NvInfer.h:2869
void setNbOutputMaps(int32_t nbOutputMaps) noexcept
Set the number of output maps for the convolution.
Definition: NvInfer.h:1029
bool precisionIsSet() const noexcept
whether the computational precision has been set for this layer
Definition: NvInfer.h:649
bool setWeightsName(Weights weights, const char *name) noexcept
Associate a name with all current uses of the given weights.
Definition: NvInfer.h:7105
The TensorRT API version 1 namespace.
constexpr int32_t EnumMax< RNNGateType >() noexcept
Maximum number of elements in RNNGateType enum.
Definition: NvInfer.h:3219
TRT_DEPRECATED void setPadding(DimsHW padding) noexcept
Set the padding of the convolution.
Definition: NvInfer.h:1085
Base class for all layer classes in a network definition.
Definition: NvInfer.h:517
ITensor * getCellState() const noexcept
Get the initial cell state of the RNN.
Definition: NvInfer.h:3497
constexpr int32_t EnumMax< ReduceOperation >() noexcept
Maximum number of elements in ReduceOperation enum.
Definition: NvInfer.h:3639
TensorFormats getAllowedFormats() const noexcept
Get a bitmask of TensorFormat values that the tensor supports. For a shape tensor,...
Definition: NvInfer.h:443
void setStrideNd(Dims stride) noexcept
Set the multi-dimension stride of the convolution.
Definition: NvInfer.h:1322
void setWindowSizeNd(Dims windowSize) noexcept
Set the multi-dimension window size for pooling.
Definition: NvInfer.h:1917
void setStart(Dims start) noexcept
Set the start offset that the slice layer uses to create the output slice.
Definition: NvInfer.h:4089
Inverse hyperbolic tangent.
void setCoordinateTransformation(ResizeCoordinateTransformation coordTransform) noexcept
Set coordinate transformation function.
Definition: NvInfer.h:4856
void setKernelSizeNd(Dims kernelSize) noexcept
Set the multi-dimension kernel size of the convolution.
Definition: NvInfer.h:1297
void setReduceAxes(uint32_t reduceAxes) noexcept
Set the axes over which to reduce.
Definition: NvInfer.h:3679
IReduceLayer * addReduce(ITensor &input, ReduceOperation operation, uint32_t reduceAxes, bool keepDimensions) noexcept
Add a reduce layer to the network.
Definition: NvInfer.h:6446
Definition: NvInfer.h:5286
Thresholded ReLU activation: x>alpha ? x : 0.
void setBiasWeights(Weights weights) noexcept
Set the bias weights for the deconvolution.
Definition: NvInfer.h:2562
Layer type for pluginV2.
Definition: NvInfer.h:3516
void setMode(SliceMode mode) noexcept
Set the slice mode.
Definition: NvInfer.h:4172
void setOperation(ReduceOperation op) noexcept
Set the reduce operation for the layer.
Definition: NvInfer.h:3659
Enables strict type constraints.
void setPrePadding(Dims padding) noexcept
Set the multi-dimension pre-padding for pooling.
Definition: NvInfer.h:1842
IConcatenationLayer * addConcatenation(ITensor *const *inputs, int32_t nbInputs) noexcept
Add a concatenation layer to the network.
Definition: NvInfer.h:6222
constexpr int32_t EnumMax< LoopOutput >() noexcept
Maximum number of elements in LoopOutput enum.
Definition: NvInfer.h:4944
bool platformHasFastFp16() const noexcept
Determine whether the platform has fast native fp16.
Definition: NvInfer.h:8513
int32_t getNbOutputMaps() const noexcept
Get the number of output feature maps for the deconvolution.
Definition: NvInfer.h:2430
nvinfer1::IHostMemory * serialize() const noexcept
Serialize a timing cache to IHostMemory object.
Definition: NvInfer.h:7782
int32_t getHiddenSize() const noexcept
Get the hidden size of the RNN.
Definition: NvInfer.h:3242
IResizeLayer * addResize(ITensor &input) noexcept
Add a resize layer to the network.
Definition: NvInfer.h:6962
Dims getPrePaddingNd() const noexcept
Get the padding that is applied at the start of the tensor.
Definition: NvInfer.h:3805
A elementwise layer in a network definition.
Definition: NvInfer.h:2845
bool setTacticSources(TacticSources tacticSources) noexcept
Set tactic sources.
Definition: NvInfer.h:8349
Network iterations from first input to last input.
MatrixOperation getOperation(int32_t index) const noexcept
Get the operation for an input tensor.
Definition: NvInfer.h:4408
Inverse hyperbolic cosine.
Weights getBiasWeights() const noexcept
Get the bias weights for the convolution.
Definition: NvInfer.h:1176
void setErrorRecorder(IErrorRecorder *recorder) noexcept
Set the ErrorRecorder for this interface.
Definition: NvInfer.h:7124
void setPostPadding(Dims padding) noexcept
Set the multi-dimension post-padding of the deconvolution.
Definition: NvInfer.h:2618
void setAxis(int32_t axis) noexcept
Set the axis used by ScatterMode::kELEMENTS.
Definition: NvInfer.h:5972
Tensor is a scalar of type kBOOL. Loop terminates when value is false.
const char * getName() const noexcept
Return the name of a layer.
Definition: NvInfer.h:548
RNNGateType
Identifies an individual gate within an RNN cell.
Definition: NvInfer.h:3207
void setOutputType(int32_t index, DataType dataType) noexcept
Set the output type of this layer.
Definition: NvInfer.h:690
ILoop * getLoop() const noexcept
Return pointer to ILoop associated with this boundary layer.
Definition: NvInfer.h:4970
void setKernelWeights(Weights weights) noexcept
Set the kernel weights for the convolution.
Definition: NvInfer.h:1141
Check if two elements are equal.
void setNbOutputChannels(int32_t nbOutputs) noexcept
Set the number of output channels K from the fully connected layer.
Definition: NvInfer.h:1458
Weights getKernelWeights() const noexcept
Get the kernel weights.
Definition: NvInfer.h:1488
constexpr int32_t EnumMax< CalibrationAlgoType >() noexcept
Maximum number of elements in CalibrationAlgoType enum.
Definition: NvInfer.h:7250
void setK(int32_t k) noexcept
Set the k value for the layer.
Definition: NvInfer.h:4289
Dims getDimensions() const noexcept
Get the dimensions of a tensor.
Definition: NvInfer.h:252
ISliceLayer * addSlice(ITensor &input, Dims start, Dims size, Dims stride) noexcept
Add a slice layer to the network.
Definition: NvInfer.h:6725
int32_t getDLACore() const noexcept
Get the DLA core that the engine executes on.
Definition: NvInfer.h:8095
ResizeCoordinateTransformation getCoordinateTransformation() const noexcept
Get coordinate transformation function.
Definition: NvInfer.h:4866
void setNearestRounding(ResizeRoundMode value) noexcept
Set rounding mode for nearest neighbor resize.
Definition: NvInfer.h:4909
EngineCapability getEngineCapability() const noexcept
Query EngineCapability flow configured for the builder.
Definition: NvInfer.h:7907
void setNbGroups(int32_t nbGroups) noexcept
Set the number of groups for a deconvolution.
Definition: NvInfer.h:2513
ITripLimitLayer * addTripLimit(ITensor &tensor, TripLimit limit) noexcept
Add a trip-count limiter, based on the given tensor.
Definition: NvInfer.h:5316
IFillLayer * addFill(Dims dimensions, FillOperation op) noexcept
Add a fill layer to the network.
Definition: NvInfer.h:7068
void setName(const char *name) noexcept
Set the name of a layer.
Definition: NvInfer.h:537
void setOperation(UnaryOperation op) noexcept
Set the unary operation for the layer.
Definition: NvInfer.h:3590
UnaryOperation
Enumerates the unary operations that may be performed by a Unary layer.
Definition: NvInfer.h:3541
Disable reuse of timing information across identical layers.
Class to handle tactic timing info collected from builder.
Definition: NvInfer.h:7768
Output value is value of tensor for last iteration.
int32_t getAxis() const noexcept
Get the axis.
Definition: NvInfer.h:5980
int32_t getWindowSize() const noexcept
Get the LRN window size.
Definition: NvInfer.h:2024
Dims getPrePadding() const noexcept
Get the pre-padding.
Definition: NvInfer.h:1852
ITensor * getInput(int32_t index) const noexcept
Get the layer input corresponding to the given index.
Definition: NvInfer.h:569
ResizeRoundMode getNearestRounding() const noexcept
Get rounding mode for nearest neighbor resize.
Definition: NvInfer.h:4919
Dims getSize() const noexcept
Get dimensions of the output slice.
Definition: NvInfer.h:4133
constexpr int32_t EnumMax< RNNOperation >() noexcept
Maximum number of elements in RNNOperation enum.
Definition: NvInfer.h:3147
constexpr int32_t EnumMax< ScatterMode >() noexcept
Maximum number of elements in ScatterMode enum.
Definition: NvInfer.h:5883
void setPower(Weights power) noexcept
Set the power value.
Definition: NvInfer.h:2210
Layer type for getting shape of a tensor.
Definition: NvInfer.h:4227
Coordinates wrap around periodically.
const char * getName() const noexcept
Get the tensor name.
Definition: NvInfer.h:220
ITopKLayer * addTopK(ITensor &input, TopKOperation op, int32_t k, uint32_t reduceAxes) noexcept
Add a TopK layer to the network.
Definition: NvInfer.h:6480
void setAxis(int32_t axis) noexcept
Set the quantization axis.
Definition: NvInfer.h:5792
float getBeta() const noexcept
Get the LRN beta value.
Definition: NvInfer.h:2066
void setInput(int32_t index, ITensor &tensor) noexcept
Replace an input of this layer with a specific tensor.
Definition: NvInfer.h:605
IEinsumLayer * addEinsum(ITensor *const *inputs, int32_t nbInputs, const char *equation) noexcept
Add an Einsum layer to the network.
Definition: NvInfer.h:7226
void setMessage(const char *message) noexcept
Set the message to print if the assertion fails.
Definition: NvInfer.h:5409
NetworkDefinitionCreationFlag
List of immutable network properties expressed at network creation time. NetworkDefinitionCreationFla...
Definition: NvInfer.h:8442
std::size_t getMaxWorkspaceSize() const noexcept
Get the maximum workspace size.
Definition: NvInfer.h:7951
ND (0 < N <= 8) nearest neighbor resizing.
const char * getName() const noexcept
Return the name of the loop.
Definition: NvInfer.h:5364
void setShift(Weights shift) noexcept
Set the shift value.
Definition: NvInfer.h:2170
Generate evenly spaced numbers over a specified interval.
PaddingMode
Enumerates the modes of padding to perform in convolution, deconvolution and pooling layer,...
Definition: NvInfer.h:961
void setAlpha(double alpha) noexcept
Set the alpha parameter.
Definition: NvInfer.h:5540
TRT_DEPRECATED DimsHW getPostPadding() const noexcept
Get the padding that is applied at the end of the tensor.
Definition: NvInfer.h:3779
void setEngineCapability(EngineCapability capability) noexcept
Configure the builder to target specified EngineCapability flow.
Definition: NvInfer.h:7895
void setName(const char *name) noexcept
Set the name of the conditional.
Definition: NvInfer.h:5108
DeviceType getDeviceType(const ILayer *layer) const noexcept
Get the device that this layer executes on.
Definition: NvInfer.h:8040
Perform the normal matrix multiplication in the first recurrent layer.
Dims getPostPadding() const noexcept
Get the padding.
Definition: NvInfer.h:1880
CalibrationAlgoType getAlgorithm() noexcept override
Definition: NvInfer.h:7390
void setMode(ScaleMode mode) noexcept
Set the scale mode.
Definition: NvInfer.h:2150
void setMode(ScatterMode mode) noexcept
Set the scatter mode.
Definition: NvInfer.h:5952
PaddingMode getPaddingMode() const noexcept
Get the padding mode.
Definition: NvInfer.h:1284
IActivationLayer * addActivation(ITensor &input, ActivationType type) noexcept
Add an activation layer to the network.
Definition: NvInfer.h:6127
bool reset() noexcept
Empty the timing cache.
Definition: NvInfer.h:7816
MatrixOperation
Enumerates the operations that may be performed on a tensor by IMatrixMultiplyLayer before multiplica...
Definition: NvInfer.h:4335
ITensor * getOutput(int32_t index) const noexcept
Get the layer output corresponding to the given index.
Definition: NvInfer.h:588
const char * getName() const noexcept
Return name of the algorithm node. This is a unique identifier for the IAlgorithmContext.
Definition: NvInfer.h:7530
Use formula to map the original index.
Optimization profile for dynamic input dimensions and shape tensors.
Definition: NvInferRuntime.h:1092
TRT_DEPRECATED IPaddingLayer * addPaddingNd(ITensor &input, Dims prePadding, Dims postPadding) noexcept
Add a padding layer to the network. Only 2D padding is currently supported.
Definition: NvInfer.h:7085
PoolingType
The type of pooling to perform in a pooling layer.
Definition: NvInfer.h:1636
float getDynamicRangeMax() const noexcept
Get maximum of dynamic range.
Definition: NvInfer.h:411
ILRNLayer * addLRN(ITensor &input, int32_t window, float alpha, float beta, float k) noexcept
Add a LRN layer to the network.
Definition: NvInfer.h:6165
void setBeta(float beta) noexcept
Set the LRN beta value.
Definition: NvInfer.h:2056
void resetOutputType(int32_t index) noexcept
reset the output type for this layer
Definition: NvInfer.h:729
DataType getType() const noexcept
Get the data type of a tensor.
Definition: NvInfer.h:279
ActivationType getActivationType() const noexcept
Get the type of activation to be performed.
Definition: NvInfer.h:1574
ResizeSelector
The coordinate selector when resize to single pixel output.
Definition: NvInfer.h:4613
Output value is concatenation of values of tensor for each iteration, in forward order.
void setHiddenState(ITensor &hidden) noexcept
Set the initial hidden state of the RNN with the provided hidden ITensor.
Definition: NvInfer.h:3460
DataType getPrecision() const noexcept
get the computational precision of this layer
Definition: NvInfer.h:637
ITensor * getSequenceLengths() const noexcept
Get the sequence lengths specified for the RNN.
Definition: NvInfer.h:3281
uint32_t TacticSources
Represents a collection of one or more TacticSource values combine using bitwise-OR operations.
Definition: NvInferRuntime.h:1274
DataType
The type of weights and tensors.
Definition: NvInferRuntimeCommon.h:150
void setMaxBatchSize(int32_t batchSize) noexcept
Set the maximum batch size.
Definition: NvInfer.h:8492
Dims getKernelSizeNd() const noexcept
Get the multi-dimension kernel size of the deconvolution.
Definition: NvInfer.h:2679
void setQuantizationFlags(QuantizationFlags flags) noexcept
Set the quantization flags.
Definition: NvInfer.h:8276
Elu activation: x>=0 ? x : alpha * (exp(x) - 1).
int32_t getMaxBatchSize() const noexcept
Get the maximum batch size.
Definition: NvInfer.h:8505
void resetDeviceType(const ILayer *layer) noexcept
reset the DeviceType for this layer
Definition: NvInfer.h:8060
A concatenation layer in a network definition.
Definition: NvInfer.h:2342
Layer type for shuffling data.
Definition: NvInfer.h:3864
TRT_DEPRECATED DimsHW getPrePadding() const noexcept
Get the padding that is applied at the start of the tensor.
Definition: NvInfer.h:3753
bool getZeroIsPlaceholder() const noexcept
Get meaning of 0 in reshape dimensions.
Definition: NvInfer.h:4014
void setMode(GatherMode mode) noexcept
Set the gather mode.
Definition: NvInfer.h:3038
Weights getShift() const noexcept
Get the shift value.
Definition: NvInfer.h:2180
RNNOperation
Enumerates the RNN operations that may be performed by an RNN layer.
Definition: NvInfer.h:3137
A tensor in a network definition.
Definition: NvInfer.h:193
Dims getKernelSizeNd() const noexcept
Get the multi-dimension kernel size of the convolution.
Definition: NvInfer.h:1307
Definition: NvInfer.h:5129
BuilderFlags getFlags() const noexcept
Get the build mode flags for this builder config. Defaults to 0.
Definition: NvInfer.h:7980
ILoop * addLoop() noexcept
Add a loop to the network.
Definition: NvInfer.h:6995
uint32_t QuantizationFlags
Represents one or more QuantizationFlag values using binary OR operations.
Definition: NvInfer.h:7686
Select the upper left pixel.
int32_t getDataLength() const noexcept
Get the maximum data length of the RNN.
Definition: NvInfer.h:3250
PaddingMode getPaddingMode() const noexcept
Get the padding mode.
Definition: NvInfer.h:1904
Floor division of the first element by the second.
void setGatherAxis(int32_t axis) noexcept
Set the axis used by GatherMode::kELEMENTS and GatherMode::kDEFAULT The axis must be less than the nu...
Definition: NvInfer.h:2988
const char * getMessage() const noexcept
Return the assertion message.
Definition: NvInfer.h:5419
LayerType
The type values of layer classes.
Definition: NvInfer.h:89
int32_t getNbElementWiseDims() const noexcept
Get the number of leading dimensions of indices tensor to be handled elementwise.
Definition: NvInfer.h:3028
void setProfileStream(const cudaStream_t stream) noexcept
Set the cuda stream that is used to profile this network.
Definition: NvInfer.h:8151
nvinfer1::IBuilderConfig * createBuilderConfig() noexcept
Create a builder configuration object.
Definition: NvInfer.h:8579
void resetPrecision() noexcept
reset the computational precision for this layer
Definition: NvInfer.h:659
Identical coefficients across all elements of the tensor.
A LRN layer in a network definition.
Definition: NvInfer.h:2002
RNNDirection
Enumerates the RNN direction that may be performed by an RNN layer.
Definition: NvInfer.h:3159
DeviceType getDefaultDeviceType() const noexcept
Get the default DeviceType which was set by setDefaultDeviceType.
Definition: NvInfer.h:8115
Four-gate LSTM network w/o peephole connections.
ILayer * getLayer(int32_t index) const noexcept
Get the layer specified by the given index.
Definition: NvInfer.h:6348
Check if element in first tensor is greater than corresponding element in second tensor.
FillOperation getOperation() const noexcept
Get the fill operation for the layer.
Definition: NvInfer.h:5522
void reset() noexcept
Resets the builder state to default values.
Definition: NvInfer.h:8667
TripLimit
Enum that describes kinds of trip limits.
Definition: NvInfer.h:4950
Layer that represents a reduction operator across Shape, Int32, Float, Half, and Int8 tensors.
Definition: NvInfer.h:3651
ResizeMode getResizeMode() const noexcept
Get resize mode for an input tensor.
Definition: NvInfer.h:4789
Layer that represents a Matrix Multiplication.
Definition: NvInfer.h:4389
float getBlendFactor() const noexcept
Get the blending factor for the max_average_blend mode: max_average_blendPool = (1-blendFactor)*maxPo...
Definition: NvInfer.h:1796
Logical XOR of two elements.
int32_t getAxis() const noexcept
Get the quantization axis.
Definition: NvInfer.h:5781
Substract the second element from the first.
void setDimensions(Dims dimensions) noexcept
Set the dimensions for the layer.
Definition: NvInfer.h:4499
LeakyRelu activation: x>=0 ? x : alpha * x.
FillOperation
Enumerates the tensor fill operations that may performed by a fill layer.
Definition: NvInfer.h:5437
IScatterLayer * addScatter(ITensor &data, ITensor &indices, ITensor &updates, ScatterMode mode) noexcept
Add a Scatter layer to the network with specified mode and axis=0.
Definition: NvInfer.h:7178
TRT_DEPRECATED IPoolingLayer * addPooling(ITensor &input, PoolingType type, DimsHW windowSize) noexcept
Add a pooling layer to the network.
Definition: NvInfer.h:6146
Forward declaration of IEngineInspector for use by other interfaces.
Definition: NvInferRuntime.h:79
void setSize(Dims size) noexcept
Set the dimensions of the output slice.
Definition: NvInfer.h:4118
void setBiasForGate(int32_t layerIndex, RNNGateType gate, bool isW, Weights bias) noexcept
Set the bias parameters for an individual gate in the RNN.
Definition: NvInfer.h:3434
Definition: NvInfer.h:7349
TRT_DEPRECATED bool hasExplicitPrecision() const noexcept
True if network is an explicit precision network.
Definition: NvInfer.h:6979
Generate an output tensor with specified mode.
Definition: NvInfer.h:5475
QuantizationFlag
List of valid flags for quantizing the network to int8.
Definition: NvInfer.h:7695
double getAlpha() const noexcept
Get the value of alpha parameter.
Definition: NvInfer.h:5555
ILoopOutputLayer * addLoopOutput(ITensor &tensor, LoopOutput outputKind, int32_t axis=0) noexcept
Make an output for this loop, based on the given tensor.
Definition: NvInfer.h:5341
void setQuantizationFlag(QuantizationFlag flag) noexcept
Set a single quantization flag.
Definition: NvInfer.h:8312
Application-implemented interface for calibration.
Definition: NvInfer.h:7266
Reference counted application-implemented error reporting interface for TensorRT objects.
Definition: NvInferRuntimeCommon.h:1439
IIfConditional * addIfConditional() noexcept
Add an If-conditional layer to the network.
Definition: NvInfer.h:7212
void setGpuAllocator(IGpuAllocator *allocator) noexcept
Set the GPU allocator.
Definition: NvInfer.h:8569
Layer that represents a TopK reduction.
Definition: NvInfer.h:4259
void setResizeMode(ResizeMode resizeMode) noexcept
Set resize mode for an input tensor.
Definition: NvInfer.h:4779
TRT_DEPRECATED void setPrePadding(DimsHW padding) noexcept
Set the padding that is applied at the start of the tensor.
Definition: NvInfer.h:3741
ResizeRoundMode
The rounding mode for nearest neighbor resize.
Definition: NvInfer.h:4639
Dims getPaddingNd() const noexcept
Get the multi-dimension padding of the convolution.
Definition: NvInfer.h:1362
int32_t getAxis() const noexcept
Get axis being iterated over.
Definition: NvInfer.h:5255
OptProfileSelector
When setting or querying optimization profile parameters (such as shape tensor inputs or dynamic dime...
Definition: NvInferRuntime.h:1056
TensorLocation getLocation() const noexcept
Get the storage location of a tensor.
Definition: NvInfer.h:358
void setPrePadding(Dims padding) noexcept
Set the multi-dimension pre-padding of the convolution.
Definition: NvInfer.h:1221
uint32_t BuilderFlags
Represents one or more QuantizationFlag values using binary OR operations, e.g., 1U << BuilderFlag::k...
Definition: NvInfer.h:7716
Use CAFFE padding, rounding output size up, uses prePadding value.
Definition: NvInfer.h:5377
void setAxis(int32_t axis) noexcept
Set the axis along which concatenation occurs.
Definition: NvInfer.h:2355
uint32_t NetworkDefinitionCreationFlags
Represents one or more NetworkDefinitionCreationFlag flags using binary OR operations....
Definition: NvInfer.h:8431
int32_t getNbInputs() const noexcept
Get the number of inputs of a layer.
Definition: NvInfer.h:556
Weights getWeights() const noexcept
Get the weights for the layer.
Definition: NvInfer.h:4487
IGatherLayer * addGather(ITensor &data, ITensor &indices, int32_t axis) noexcept
Add gather with mode GatherMode::kDEFAULT and specified axis and nbElementWiseDims=0.
Definition: NvInfer.h:6496
Dims getDilationNd() const noexcept
Get the multi-dimension dilation of the deconvolution.
Definition: NvInfer.h:2781
ITensor * getOutput(int32_t index) const noexcept
Get the output tensor specified by the given index.
Definition: NvInfer.h:6406
IIdentityLayer * addIdentity(ITensor &input) noexcept
Add an identity layer.
Definition: NvInfer.h:6660
TRT_DEPRECATED void setStride(DimsHW stride) noexcept
Set the stride of the deconvolution.
Definition: NvInfer.h:2447
bool dynamicRangeIsSet() const noexcept
Query whether dynamic range is set.
Definition: NvInfer.h:383
ScatterMode
Control form of IScatterLayer.
Definition: NvInfer.h:5875
Definition: NvInfer.h:7331
constexpr int32_t EnumMax< UnaryOperation >() noexcept
Maximum number of elements in UnaryOperation enum.
Definition: NvInfer.h:3570
A Gather layer in a network definition. Supports several kinds of gathering.
Definition: NvInfer.h:2976
Dims getPrePadding() const noexcept
Get the pre-padding.
Definition: NvInfer.h:2600
Dims getDimensions() const noexcept
Get the dimensions for the layer.
Definition: NvInfer.h:4511
void setOperation(FillOperation op) noexcept
Set the fill operation for the layer.
Definition: NvInfer.h:5512
virtual int32_t getMinTimingIterations() const noexcept
Query the number of minimization iterations.
Definition: NvInfer.h:7857
void setAxis(int32_t axis) noexcept
Set axis to iterate over.
Definition: NvInfer.h:5249
Dims getPaddingNd() const noexcept
Get the multi-dimension padding of the deconvolution.
Definition: NvInfer.h:2737
uint32_t getReduceAxes() const noexcept
Get the axes over which to reduce for the layer.
Definition: NvInfer.h:3689
static constexpr int32_t MAX_DIMS
The maximum rank (number of dimensions) supported for a tensor.
Definition: NvInferRuntimeCommon.h:193
void setOperation(ElementWiseOperation op) noexcept
Set the binary operation for the layer.
Definition: NvInfer.h:2857
virtual void setAvgTimingIterations(int32_t avgTiming) noexcept
Set the number of averaging iterations used when timing layers.
Definition: NvInfer.h:7870
void setName(const char *name) noexcept
Set the tensor name.
Definition: NvInfer.h:208
Dims getPaddingNd() const noexcept
Get the multi-dimension padding for pooling.
Definition: NvInfer.h:1983
Definition: NvInfer.h:5055
virtual void setMinTimingIterations(int32_t minTiming) noexcept
Set the number of minimization iterations used when timing layers.
Definition: NvInfer.h:7845
constexpr int32_t EnumMax< LayerType >() noexcept
Maximum number of elements in LayerType enum.
Definition: NvInfer.h:136
CalibrationAlgoType getAlgorithm() noexcept override
Definition: NvInfer.h:7372
Product of the two elements.
double getBeta() const noexcept
Get the value of beta parameter.
Definition: NvInfer.h:5588
void setNbGroups(int32_t nbGroups) noexcept
Set the number of groups for a convolution.
Definition: NvInfer.h:1117
Definition: NvInfer.h:4966
uint32_t TensorFormats
It is capable of representing one or more TensorFormat by binary OR operations, e....
Definition: NvInfer.h:147
void setPaddingNd(Dims padding) noexcept
Set the multi-dimension padding of the convolution.
Definition: NvInfer.h:1350
EngineCapability
List of supported engine capability flows.
Definition: NvInferRuntime.h:105
Describes a variation of execution of a layer. An algorithm is represented by IAlgorithmVariant and t...
Definition: NvInfer.h:7576
Weights getWeightsForGate(int32_t layerIndex, RNNGateType gate, bool isW) const noexcept
Get the weight parameters for an individual gate in the RNN.
Definition: NvInfer.h:3409
nvinfer1::IOptimizationProfile * createOptimizationProfile() noexcept
Create a new optimization profile.
Definition: NvInfer.h:8625
void setDLACore(int32_t dlaCore) noexcept
Sets the DLA core used by the network.
Definition: NvInfer.h:8084
void setSelectorForSinglePixel(ResizeSelector selector) noexcept
Set coordinate selector function when resized to single pixel.
Definition: NvInfer.h:4883
IAlgorithmSelector * getAlgorithmSelector() const noexcept
Get Algorithm Selector.
Definition: NvInfer.h:8234
int32_t getNbInputs() const noexcept
Get the number of inputs in the network.
Definition: NvInfer.h:6360
bool isExecutionTensor() const noexcept
Whether the tensor is an execution tensor.
Definition: NvInfer.h:500
void setName(const char *name) noexcept
Set the name of the loop.
Definition: NvInfer.h:5354
constexpr int32_t EnumMax< GatherMode >() noexcept
Maximum number of elements in GatherMode enum.
Definition: NvInfer.h:2893
void setFlags(BuilderFlags builderFlags) noexcept
Set the build mode flags to turn on builder options for this network.
Definition: NvInfer.h:7968
ResizeCoordinateTransformation
The resize coordinate transformation function.
Definition: NvInfer.h:4563
TRT_DEPRECATED DimsHW getStride() const noexcept
Get the stride of the convolution.
Definition: NvInfer.h:1065
const nvinfer1::ITimingCache * getTimingCache() const noexcept
Get the pointer to the timing cache from current IBuilderConfig.
Definition: NvInfer.h:8416
void setSecondTranspose(Permutation permutation) noexcept
Set the permutation applied by the second transpose operation.
Definition: NvInfer.h:3973
IScaleLayer * addScaleNd(ITensor &input, ScaleMode mode, Weights shift, Weights scale, Weights power, int32_t channelAxis) noexcept
Add a multi-dimension scale layer to the network.
Definition: NvInfer.h:6946
Similar to ONNX GatherElements.
IAssertionLayer * addAssertion(ITensor &condition, const char *message) noexcept
Add an assertion layer to the network.
Definition: NvInfer.h:7050
void setDeviceType(const ILayer *layer, DeviceType deviceType) noexcept
Set the device that this layer must execute on.
Definition: NvInfer.h:8031
Conditional Output layer.
Definition: NvInfer.h:7384
IParametricReLULayer * addParametricReLU(ITensor &input, ITensor &slope) noexcept
Add a parametric ReLU layer to the network.
Definition: NvInfer.h:6846
bool getAverageCountExcludesPadding() const noexcept
Get whether average pooling uses as a denominator the overlap area between the window and the unpadde...
Definition: NvInfer.h:1824
Check if element in first tensor is less than corresponding element in second tensor.
TRT_DEPRECATED IConvolutionLayer * addConvolution(ITensor &input, int32_t nbOutputMaps, DimsHW kernelSize, Weights kernelWeights, Weights biasWeights) noexcept
Add a convolution layer to the network.
Definition: NvInfer.h:6086
TRT_DEPRECATED void setDilation(DimsHW dilation) noexcept
Set the dilation for a convolution.
Definition: NvInfer.h:1192
Weights getPower() const noexcept
Get the power value.
Definition: NvInfer.h:2220
int32_t getAxis() const noexcept
Get the quantization axis.
Definition: NvInfer.h:5694
IPluginV2 & getPlugin() noexcept
Get the plugin for the layer.
Definition: NvInfer.h:3524
Definition: NvInfer.h:5015
Weights getKernelWeights() const noexcept
Get the kernel weights of the convolution.
Definition: NvInfer.h:1151
A Scale layer in a network definition.
Definition: NvInfer.h:2142
const char * getName() const noexcept
Returns the name associated with the network.
Definition: NvInfer.h:6761
TRT_DEPRECATED void setPadding(DimsHW padding) noexcept
Set the padding of the deconvolution.
Definition: NvInfer.h:2479
void setWeights(Weights weights) noexcept
Set the weights for the layer.
Definition: NvInfer.h:4477
Three-gate network consisting of Gated Recurrent Units.
GatherMode
Control form of IGatherLayer.
Definition: NvInfer.h:2884
Dims getStride() const noexcept
Get the stride for the output slice.
Definition: NvInfer.h:4162
void setAlpha(float alpha) noexcept
Set the LRN alpha value.
Definition: NvInfer.h:2035
TRT_DEPRECATED nvinfer1::ICudaEngine * buildEngineWithConfig(INetworkDefinition &network, IBuilderConfig &config) noexcept
Builds an engine for the given INetworkDefinition and given IBuilderConfig.
Definition: NvInfer.h:8594
A Softmax layer in a network definition.
Definition: NvInfer.h:2277
TRT_DEPRECATED DimsHW getStride() const noexcept
Get the stride for pooling.
Definition: NvInfer.h:1738
IPoolingLayer * addPoolingNd(ITensor &input, PoolingType type, Dims windowSize) noexcept
Add a multi-dimension pooling layer to the network.
Definition: NvInfer.h:6888
void setName(const char *name) noexcept
Sets the name of the network.
Definition: NvInfer.h:6747
Tensor is scalar of type kINT32 that contains the trip count.
bool platformHasTf32() const noexcept
Determine whether the platform has TF32 support.
Definition: NvInfer.h:8675
int32_t getMaxSeqLength() const noexcept
Get the maximum sequence length of the RNN.
Definition: NvInfer.h:3246
Weights getBiasWeights() const noexcept
Get the bias weights.
Definition: NvInfer.h:1510
ILogger * getLogger() const noexcept
get the logger with which the builder was created
Definition: NvInfer.h:8726
Holds properties for configuring a builder to produce an engine.
Definition: NvInfer.h:7830
void setPaddingMode(PaddingMode paddingMode) noexcept
Set the padding mode.
Definition: NvInfer.h:2642
bool setCalibrationProfile(const IOptimizationProfile *profile) noexcept
Add a calibration profile.
Definition: NvInfer.h:8249
TRT_DEPRECATED void setKernelSize(DimsHW kernelSize) noexcept
Set the HW kernel size of the convolution.
Definition: NvInfer.h:2396
ITensor * getInput(int32_t index) const noexcept
Get the input tensor specified by the given index.
Definition: NvInfer.h:6376
TRT_DEPRECATED DimsHW getStride() const noexcept
Get the stride of the deconvolution.
Definition: NvInfer.h:2459
ITensor * getHiddenState() const noexcept
Get the initial hidden state of the RNN.
Definition: NvInfer.h:3469
void setAverageCountExcludesPadding(bool exclusive) noexcept
Set whether average pooling uses as a denominator the overlap area between the window and the unpadde...
Definition: NvInfer.h:1813
IGatherLayer * addGatherV2(ITensor &data, ITensor &indices, GatherMode mode)
Add gather with specified mode, axis=0 and nbElementWiseDims=0.
Definition: NvInfer.h:6512
An Einsum layer in a network.
Definition: NvInfer.h:5838
constexpr int32_t EnumMax< NetworkDefinitionCreationFlag >() noexcept
Maximum number of elements in NetworkDefinitionCreationFlag enum.
Definition: NvInfer.h:8467
constexpr int32_t EnumMax< TripLimit >() noexcept
Maximum number of elements in TripLimit enum.
Definition: NvInfer.h:4959
Weights getBiasForGate(int32_t layerIndex, RNNGateType gate, bool isW) const noexcept
Get the bias parameters for an individual gate in the RNN.
Definition: NvInfer.h:3443
A resize layer in a network definition.
Definition: NvInfer.h:4685
IMatrixMultiplyLayer * addMatrixMultiply(ITensor &input0, MatrixOperation op0, ITensor &input1, MatrixOperation op1) noexcept
Add a MatrixMultiply layer to the network.
Definition: NvInfer.h:6549
Enable debugging of layers via synchronizing after every layer.
const TRT_DEPRECATED IAlgorithmIOInfo & getAlgorithmIOInfo(int32_t index) const noexcept
Returns the format of an Algorithm input or output. Algorithm inputs are incrementally numbered first...
Definition: NvInfer.h:7589
ProfilingVerbosity getProfilingVerbosity() const noexcept
Get verbosity level of layer information exposed in NVTX annotations and IEngineInspector.
Definition: NvInfer.h:8217
void setKernelSizeNd(Dims kernelSize) noexcept
Set the multi-dimension kernel size of the deconvolution.
Definition: NvInfer.h:2669
bool isDeviceTypeSet(const ILayer *layer) const noexcept
whether the DeviceType has been explicitly set for this layer
Definition: NvInfer.h:8050
bool isNetworkInput() const noexcept
Whether the tensor is a network input.
Definition: NvInfer.h:302
void setScales(const float *scales, int32_t nbScales) noexcept
Set the resize scales.
Definition: NvInfer.h:4746
LayerType getType() const noexcept
Return the type of a layer.
Definition: NvInfer.h:525
const char * getName() const noexcept
Return the name of the conditional.
Definition: NvInfer.h:5118
A Pooling layer in a network definition.
Definition: NvInfer.h:1664
void setPaddingMode(PaddingMode paddingMode) noexcept
Set the padding mode.
Definition: NvInfer.h:1272
CalibrationAlgoType getAlgorithm() noexcept override
Definition: NvInfer.h:7355
uint32_t getReduceAxes() const noexcept
Get the axes to reduce for the layer.
Definition: NvInfer.h:4319
nvinfer1::ITimingCache * createTimingCache(const void *blob, std::size_t size) const noexcept
Create timing cache.
Definition: NvInfer.h:8383
void setCellState(ITensor &cell) noexcept
Set the initial cell state of the LSTM with the provided cell ITensor.
Definition: NvInfer.h:3488
Application-implemented class for controlling allocation on the GPU.
Definition: NvInferRuntimeCommon.h:1138
void setStrideNd(Dims stride) noexcept
Set the multi-dimension stride for pooling.
Definition: NvInfer.h:1942
IBuilder * createInferBuilder(ILogger &logger) noexcept
Create an instance of an IBuilder class.
Definition: NvInfer.h:8755
Layer that represents an unary operation.
Definition: NvInfer.h:3582
int64_t getTactic() const noexcept
Return tactic of the algorithm.
Definition: NvInfer.h:7505
Definition: NvInfer.h:5002
ReduceOperation getOperation() const noexcept
Get the reduce operation for the layer.
Definition: NvInfer.h:3669
TRT_DEPRECATED void setPadding(DimsHW padding) noexcept
Set the padding for pooling.
Definition: NvInfer.h:1754
SliceMode getMode() const noexcept
Get the slice mode.
Definition: NvInfer.h:4182
Round to nearest even for float datatype.
Builds an engine from a network definition.
Definition: NvInfer.h:8479
TRT_DEPRECATED void setPostPadding(DimsHW padding) noexcept
Set the padding that is applied at the end of the tensor.
Definition: NvInfer.h:3767
void setInt8Calibrator(IInt8Calibrator *calibrator) noexcept
Set Int8 Calibration interface.
Definition: NvInfer.h:7917
void setBeta(double beta) noexcept
Set the beta parameter.
Definition: NvInfer.h:5573
bool isNetworkSupported(INetworkDefinition const &network, IBuilderConfig const &config) const noexcept
Checks that a network is within the scope of the IBuilderConfig settings.
Definition: NvInfer.h:8716
Allow the builder to examine weights and use optimized functions when weights have suitable sparsity.
void setOperation(TopKOperation op) noexcept
Set the operation for the layer.
Definition: NvInfer.h:4267
An assertion layer in a network.
Definition: NvInfer.h:5398
const IAlgorithmVariant & getAlgorithmVariant() const noexcept
Returns the algorithm variant.
Definition: NvInfer.h:7597
Mark the network to be an explicit batch network.
TRT_DEPRECATED DimsHW getKernelSize() const noexcept
Get the HW kernel size of the deconvolution.
Definition: NvInfer.h:2408
A Quantize layer in a network definition.
Definition: NvInfer.h:5683
IShapeLayer * addShape(ITensor &input) noexcept
Add a shape layer to the network.
Definition: NvInfer.h:6779
TopKOperation
Enumerates the operations that may be performed by a TopK layer.
Definition: NvInfer.h:4239
int64_t getImplementation() const noexcept
Return implementation of the algorithm.
Definition: NvInfer.h:7497
IIteratorLayer * addIterator(ITensor &tensor, int32_t axis=0, bool reverse=false) noexcept
Return layer that subscripts tensor by loop iteration.
Definition: NvInfer.h:5329
int32_t getLayerCount() const noexcept
Get the layer count of the RNN.
Definition: NvInfer.h:3238
TensorLocation
The location for tensor data storage, device or host.
Definition: NvInferRuntime.h:245
int32_t getK() const noexcept
Get the k value for the layer.
Definition: NvInfer.h:4299
float getBeta() const noexcept
Get the beta parameter.
Definition: NvInfer.h:1621
int32_t getNbDLACores() const noexcept
Return the number of DLA engines available to this builder.
Definition: NvInfer.h:8553
ResizeSelector getSelectorForSinglePixel() const noexcept
Get the coordinate selector function when resized to single pixel.
Definition: NvInfer.h:4893
void setMaxWorkspaceSize(std::size_t workspaceSize) noexcept
Set the maximum workspace size.
Definition: NvInfer.h:7937
int32_t getNbOutputMaps() const noexcept
Get the number of output maps for the convolution.
Definition: NvInfer.h:1039
int32_t getMaxDLABatchSize() const noexcept
Get the maximum batch size DLA can support. For any tensor the total volume of index dimensions combi...
Definition: NvInfer.h:8545
RNNInputMode getInputMode() const noexcept
Get the input mode of the RNN layer.
Definition: NvInfer.h:3317
int32_t getNbOutputs() const noexcept
Return number of outputs of the algorithm.
Definition: NvInfer.h:7557
QuantizationFlags getQuantizationFlags() const noexcept
Get the quantization flags.
Definition: NvInfer.h:8288
bool setEquation(const char *equation) noexcept
Set the equation. The equation is a comma-separated list of subscript labels, where each label refers...
Definition: NvInfer.h:5850
int32_t getAxis() const noexcept
Get the axis along which concatenation occurs.
Definition: NvInfer.h:2365
IScaleLayer * addScale(ITensor &input, ScaleMode mode, Weights shift, Weights scale, Weights power) noexcept
Add a Scale layer to the network.
Definition: NvInfer.h:6192
#define TRT_DEPRECATED
< Items that are marked as deprecated will be removed in a future release.
Definition: NvInferRuntimeCommon.h:77
Definition: NvInfer.h:4985
void setStrideNd(Dims stride) noexcept
Set the multi-dimension stride of the deconvolution.
Definition: NvInfer.h:2697
bool getBroadcastAcrossBatch() const noexcept
Check if tensor is broadcast across the batch.
Definition: NvInfer.h:348
TRT_DEPRECATED IRNNv2Layer * addRNNv2(ITensor &input, int32_t layerCount, int32_t hiddenSize, int32_t maxSeqLen, RNNOperation op) noexcept
Add an layerCount deep RNN layer to the network with hiddenSize internal states that can take a batch...
Definition: NvInfer.h:6643
bool isNetworkOutput() const noexcept
Whether the tensor is a network output.
Definition: NvInfer.h:310
bool setTimingCache(const ITimingCache &cache, bool ignoreMismatch) noexcept
Attach a timing cache to IBuilderConfig.
Definition: NvInfer.h:8406
IIfConditionalOutputLayer * addOutput(ITensor &trueSubgraphOutput, ITensor &falseSubgraphOutput) noexcept
Add an If-conditional output.
Definition: NvInfer.h:5083
Weights getScale() const noexcept
Get the scale value.
Definition: NvInfer.h:2200
An Activation layer in a network definition.
Definition: NvInfer.h:1554
CalibrationAlgoType
Version of calibration algorithm to use.
Definition: NvInfer.h:7240
void setErrorRecorder(IErrorRecorder *recorder) noexcept
Set the ErrorRecorder for this interface.
Definition: NvInfer.h:8644
A Dequantize layer in a network definition.
Definition: NvInfer.h:5770
TRT_DEPRECATED void destroy() noexcept
Delete this IBuilderConfig.
Definition: NvInfer.h:8139
constexpr int32_t EnumMax< BuilderFlag >() noexcept
Maximum number of builder flags in BuilderFlag enum.
Definition: NvInfer.h:7753
void setInputMode(RNNInputMode op) noexcept
Set the input mode of the RNN layer.
Definition: NvInfer.h:3308
A deconvolution layer in a network definition.
Definition: NvInfer.h:2382
void setSequenceLengths(ITensor &seqLengths) noexcept
Specify individual sequence lengths in the batch with the ITensor pointed to by seqLengths.
Definition: NvInfer.h:3269