166 static constexpr int32_t kVALUE = 14;
204 mImpl->setName(name);
216 return mImpl->getName();
235 mImpl->setDimensions(dimensions);
249 return mImpl->getDimensions();
285 mImpl->setType(type);
300 return mImpl->getType();
308 return mImpl->isNetworkInput();
316 return mImpl->isNetworkOutput();
338 mImpl->setAllowedFormats(formats);
351 return mImpl->getAllowedFormats();
382 return mImpl->isShapeTensor();
403 return mImpl->isExecutionTensor();
429 mImpl->setDimensionName(index, name);
444 return mImpl->getDimensionName(index);
469 return mLayer->getType();
483 mLayer->setName(name);
493 return mLayer->getName();
501 return mLayer->getNbInputs();
514 return mLayer->getInput(index);
522 return mLayer->getNbOutputs();
532 return mLayer->getOutput(index);
549 return mLayer->setInput(index, tensor);
582 mLayer->setPrecision(dataType);
594 return mLayer->getPrecision();
608 return mLayer->precisionIsSet();
620 mLayer->resetPrecision();
670 mLayer->setOutputType(index, dataType);
685 return mLayer->getOutputType(index);
701 return mLayer->outputTypeIsSet(index);
715 return mLayer->resetOutputType(index);
733 mLayer->setMetadata(metadata);
746 return mLayer->getMetadata();
751 apiv::VLayer* mLayer;
928 static constexpr int32_t kVALUE = 4;
956 mImpl->setNbOutputMaps(nbOutputMaps);
966 return mImpl->getNbOutputMaps();
986 mImpl->setNbGroups(nbGroups);
996 return mImpl->getNbGroups();
1010 mImpl->setKernelWeights(weights);
1020 return mImpl->getKernelWeights();
1035 mImpl->setBiasWeights(weights);
1045 return mImpl->getBiasWeights();
1062 mImpl->setPrePadding(padding);
1072 return mImpl->getPrePadding();
1089 mImpl->setPostPadding(padding);
1099 return mImpl->getPostPadding();
1113 mImpl->setPaddingMode(paddingMode);
1125 return mImpl->getPaddingMode();
1138 mImpl->setKernelSizeNd(kernelSize);
1148 return mImpl->getKernelSizeNd();
1163 mImpl->setStrideNd(stride);
1173 return mImpl->getStrideNd();
1191 mImpl->setPaddingNd(padding);
1203 return mImpl->getPaddingNd();
1217 mImpl->setDilationNd(dilation);
1227 return mImpl->getDilationNd();
1276 mImpl->setActivationType(type);
1286 return mImpl->getActivationType();
1301 mImpl->setAlpha(alpha);
1315 mImpl->setBeta(beta);
1324 return mImpl->getAlpha();
1333 return mImpl->getBeta();
1363 static constexpr int32_t kVALUE = 3;
1390 mImpl->setPoolingType(type);
1400 return mImpl->getPoolingType();
1415 mImpl->setBlendFactor(blendFactor);
1428 return mImpl->getBlendFactor();
1442 mImpl->setAverageCountExcludesPadding(exclusive);
1453 return mImpl->getAverageCountExcludesPadding();
1471 mImpl->setPrePadding(padding);
1481 return mImpl->getPrePadding();
1499 mImpl->setPostPadding(padding);
1509 return mImpl->getPostPadding();
1522 mImpl->setPaddingMode(paddingMode);
1533 return mImpl->getPaddingMode();
1546 mImpl->setWindowSizeNd(windowSize);
1556 return mImpl->getWindowSizeNd();
1571 mImpl->setStrideNd(stride);
1581 return mImpl->getStrideNd();
1600 mImpl->setPaddingNd(padding);
1612 return mImpl->getPaddingNd();
1643 mImpl->setWindowSize(windowSize);
1653 return mImpl->getWindowSize();
1665 mImpl->setAlpha(alpha);
1675 return mImpl->getAlpha();
1687 mImpl->setBeta(beta);
1697 return mImpl->getBeta();
1719 return mImpl->getK();
1785 mImpl->setMode(mode);
1795 return mImpl->getMode();
1805 mImpl->setShift(shift);
1815 return mImpl->getShift();
1825 mImpl->setScale(scale);
1835 return mImpl->getScale();
1845 mImpl->setPower(power);
1855 return mImpl->getPower();
1870 return mImpl->getChannelAxis();
1891 mImpl->setChannelAxis(channelAxis);
1944 mImpl->setAxes(axes);
1954 return mImpl->getAxes();
1990 mImpl->setAxis(axis);
2000 return mImpl->getAxis();
2027 mImpl->setNbOutputMaps(nbOutputMaps);
2037 return mImpl->getNbOutputMaps();
2057 mImpl->setNbGroups(nbGroups);
2067 return mImpl->getNbGroups();
2081 mImpl->setKernelWeights(weights);
2091 return mImpl->getKernelWeights();
2106 mImpl->setBiasWeights(weights);
2116 return mImpl->getBiasWeights();
2133 mImpl->setPrePadding(padding);
2143 return mImpl->getPrePadding();
2160 mImpl->setPostPadding(padding);
2170 return mImpl->getPostPadding();
2184 mImpl->setPaddingMode(paddingMode);
2196 return mImpl->getPaddingMode();
2211 mImpl->setKernelSizeNd(kernelSize);
2221 return mImpl->getKernelSizeNd();
2238 mImpl->setStrideNd(stride);
2248 return mImpl->getStrideNd();
2266 mImpl->setPaddingNd(padding);
2278 return mImpl->getPaddingNd();
2304 mImpl->setDilationNd(dilation);
2314 return mImpl->getDilationNd();
2362 static constexpr int32_t kVALUE = 14;
2399 return mImpl->setOperation(op);
2411 return mImpl->getOperation();
2532 mImpl->setGatherAxis(axis);
2544 return mImpl->getGatherAxis();
2567 mImpl->setNbElementWiseDims(elementWiseDims);
2577 return mImpl->getNbElementWiseDims();
2587 mImpl->setMode(mode);
2597 return mImpl->getMode();
2626 return mImpl->getPlugin();
2653 return mImpl->getPlugin();
2736 mImpl->setOperation(op);
2746 return mImpl->getOperation();
2809 mImpl->setOperation(op);
2819 return mImpl->getOperation();
2829 mImpl->setReduceAxes(reduceAxes);
2839 return mImpl->getReduceAxes();
2849 mImpl->setKeepDimensions(keepDimensions);
2859 return mImpl->getKeepDimensions();
2893 mImpl->setPrePaddingNd(padding);
2905 return mImpl->getPrePaddingNd();
2919 mImpl->setPostPaddingNd(padding);
2931 return mImpl->getPostPaddingNd();
2981 mImpl->setFirstTranspose(permutation);
2993 return mImpl->getFirstTranspose();
3021 mImpl->setReshapeDimensions(dimensions);
3034 return mImpl->getReshapeDimensions();
3081 mImpl->setSecondTranspose(permutation);
3093 return mImpl->getSecondTranspose();
3109 return mImpl->setZeroIsPlaceholder(zeroIsPlaceholder);
3122 return mImpl->getZeroIsPlaceholder();
3233 mImpl->setStart(start);
3248 return mImpl->getStart();
3262 return mImpl->setSize(size);
3277 return mImpl->getSize();
3291 mImpl->setStride(stride);
3306 return mImpl->getStride();
3316 mImpl->setMode(mode);
3326 return mImpl->getMode();
3369 mImpl->setAxes(axes);
3384 return mImpl->getAxes();
3454 mImpl->setOperation(op);
3464 return mImpl->getOperation();
3492 return mImpl->getK();
3502 mImpl->setReduceAxes(reduceAxes);
3512 return mImpl->getReduceAxes();
3543 return mImpl->setIndicesType(type);
3555 return mImpl->getIndicesType();
3641 mImpl->setOperation(index, op);
3653 return mImpl->getOperation(index);
3697 return mImpl->setIndicesType(type);
3709 return mImpl->getIndicesType();
3806 mImpl->setToType(toType);
3817 return mImpl->getToType();
3846 mImpl->setWeights(weights);
3856 return mImpl->getWeights();
3868 mImpl->setDimensions(dimensions);
3880 return mImpl->getDimensions();
3926 static constexpr int32_t kVALUE = 3;
3980 static constexpr int32_t kVALUE = 3;
4010 static constexpr int32_t kVALUE = 2;
4046 static constexpr int32_t kVALUE = 4;
4109 return mImpl->setOutputDimensions(dimensions);
4119 return mImpl->getOutputDimensions();
4147 void setScales(
float const* scales, int32_t nbScales)
noexcept
4149 mImpl->setScales(scales, nbScales);
4166 int32_t
getScales(int32_t size,
float* scales)
const noexcept
4168 return mImpl->getScales(size, scales);
4180 mImpl->setResizeMode(interpolationMode);
4190 return mImpl->getResizeMode();
4225 mImpl->setCoordinateTransformation(coordTransform);
4235 return mImpl->getCoordinateTransformation();
4250 mImpl->setSelectorForSinglePixel(selector);
4260 return mImpl->getSelectorForSinglePixel();
4274 mImpl->setNearestRounding(value);
4284 return mImpl->getNearestRounding();
4306 mImpl->setCubicCoeff(A);
4316 return mImpl->getCubicCoeff();
4329 mImpl->setExcludeOutside(excludeFlag);
4339 return mImpl->getExcludeOutside();
4421 return mBoundary->getLoop();
4426 apiv::VLoopBoundaryLayer* mBoundary;
4444 return mBoundary->getConditional();
4449 apiv::VConditionalBoundaryLayer* mBoundary;
4533 return mImpl->setCondition(condition);
4551 return mImpl->addOutput(trueSubgraphOutput, falseSubgraphOutput);
4563 return mImpl->addInput(input);
4578 mImpl->setName(name);
4588 return mImpl->getName();
4658 return mImpl->getLoopOutput();
4675 mImpl->setAxis(axis);
4683 return mImpl->getAxis();
4732 return mImpl->getTripLimit();
4758 mImpl->setAxis(axis);
4766 return mImpl->getAxis();
4780 mImpl->setReverse(reverse);
4790 return mImpl->getReverse();
4819 return mImpl->addRecurrence(initialValue);
4840 return mImpl->addTripLimit(tensor, limit);
4853 return mImpl->addIterator(tensor, axis, reverse);
4866 return mImpl->addLoopOutput(tensor, outputKind, axis);
4881 mImpl->setName(name);
4891 return mImpl->getName();
4946 mImpl->setMessage(message);
4956 return mImpl->getMessage();
5058 mImpl->setDimensions(dimensions);
5073 return mImpl->getDimensions();
5083 mImpl->setOperation(op);
5093 return mImpl->getOperation();
5112 mImpl->setAlpha(alpha);
5127 return mImpl->getAlpha();
5146 mImpl->setBeta(beta);
5161 return mImpl->getBeta();
5222 mImpl->setAlphaInt64(alpha);
5237 return mImpl->getAlphaInt64();
5256 mImpl->setBetaInt64(beta);
5271 return mImpl->getBetaInt64();
5279 return mImpl->isAlphaBetaInt64();
5296 mImpl->setToType(toType);
5308 return mImpl->getToType();
5403 return mImpl->getAxis();
5414 mImpl->setAxis(axis);
5430 mImpl->setToType(toType);
5442 return mImpl->getToType();
5531 return mImpl->getAxis();
5542 mImpl->setAxis(axis);
5558 mImpl->setToType(toType);
5570 return mImpl->getToType();
5625 mImpl->setToType(toType);
5638 return mImpl->getToType();
5651 mImpl->setScaleType(scaleType);
5664 return mImpl->getScaleType();
5677 mImpl->setAxis(axis);
5687 return mImpl->getAxis();
5700 mImpl->setBlockSize(size);
5710 return mImpl->getBlockSize();
5766 return mImpl->setEquation(equation);
5776 return mImpl->getEquation();
5875 mImpl->setMode(mode);
5885 return mImpl->getMode();
5895 mImpl->setAxis(axis);
5903 return mImpl->getAxis();
5947 mImpl->setAxis(axis);
5955 return mImpl->getAxis();
5984 mImpl->setInterpolationMode(mode);
5996 return mImpl->getInterpolationMode();
6006 mImpl->setAlignCorners(alignCorners);
6018 return mImpl->getAlignCorners();
6030 return mImpl->setSampleMode(mode);
6042 return mImpl->getSampleMode();
6140 mImpl->setBoundingBoxFormat(fmt);
6152 return mImpl->getBoundingBoxFormat();
6166 mImpl->setTopKBoxLimit(limit);
6176 return mImpl->getTopKBoxLimit();
6211 return mImpl->setIndicesType(type);
6223 return mImpl->getIndicesType();
6256 mImpl->setBatchAxis(batchAxis);
6266 return mImpl->getBatchAxis();
6279 mImpl->setSequenceAxis(sequenceAxis);
6289 return mImpl->getSequenceAxis();
6327 return mImpl->setEpsilon(eps);
6337 return mImpl->getEpsilon();
6347 return mImpl->setAxes(axesMask);
6357 return mImpl->getAxes();
6378 return mImpl->setNbGroups(nbGroups);
6388 return mImpl->getNbGroups();
6414 return mImpl->setComputePrecision(type);
6424 return mImpl->getComputePrecision();
6519 static constexpr int32_t kVALUE = 1;
6565 return mImpl->setOperation(op);
6577 return mImpl->getOperation();
6589 mImpl->setExclusive(exclusive);
6601 return mImpl->getExclusive();
6613 mImpl->setReverse(reverse);
6625 return mImpl->getReverse();
6655 static constexpr int32_t kVALUE = 2;
6677 return mBoundary->getAttention();
6682 apiv::VAttentionBoundaryLayer* mBoundary;
6795 return mImpl->setNormalizationOperation(op);
6807 return mImpl->getNormalizationOperation();
6824 return mImpl->setMask(mask);
6836 return mImpl->getMask();
6849 return mImpl->setCausal(isCausal);
6861 return mImpl->getCausal();
6873 return mImpl->setDecomposable(decomposable);
6886 return mImpl->getDecomposable();
6905 return mImpl->setInput(index, input);
6914 return mImpl->getNbInputs();
6926 return mImpl->getInput(index);
6934 return mImpl->getNbOutputs();
6946 return mImpl->getOutput(index);
6963 return mImpl->setName(name);
6975 return mImpl->getName();
6991 return mImpl->setNormalizationQuantizeScale(tensor);
7002 return mImpl->getNormalizationQuantizeScale();
7015 return mImpl->setNormalizationQuantizeToType(type);
7027 return mImpl->getNormalizationQuantizeToType();
7095 return mImpl->addInput(name, type, dimensions);
7109 mImpl->markOutput(tensor);
7127 return mImpl->markDebug(tensor);
7143 return mImpl->unmarkDebug(tensor);
7153 return mImpl->isDebugTensor(tensor);
7175 return mImpl->markUnfusedTensorsAsDebugTensors();
7189 return mImpl->unmarkUnfusedTensorsAsDebugTensors();
7209 return mImpl->addActivation(input, type);
7228 return mImpl->addLRN(input, window, alpha, beta, k);
7254 return mImpl->addScale(input, mode, shift, scale, power);
7267 return mImpl->addSoftMax(input);
7284 return mImpl->addConcatenation(inputs, nbInputs);
7311 return mImpl->addElementWise(input1, input2, op);
7333 return mImpl->addUnary(input, operation);
7347 return mImpl->addShuffle(input);
7364 return mImpl->addOneHot(indices, values, depth, axis);
7376 return mImpl->getNbLayers();
7390 return mImpl->getLayer(index);
7402 return mImpl->getNbInputs();
7418 return mImpl->getInput(index);
7432 return mImpl->getNbOutputs();
7448 return mImpl->getOutput(index);
7475 return mImpl->addReduce(input, operation, reduceAxes, keepDimensions);
7511 return mImpl->addTopK(input, op, k, reduceAxes);
7545 return mImpl->addTopKV2(input, op, k, reduceAxes, indicesType);
7561 return mImpl->addGather(data, indices, axis);
7577 return mImpl->addGatherV2(data, indices, mode);
7596 return mImpl->addRaggedSoftMax(input, bounds);
7618 return mImpl->addMatrixMultiply(input0, op0, input1, op1);
7636 return mImpl->addNonZero(input);
7652 return mImpl->addNonZeroV2(input, indicesType);
7676 return mImpl->addConstant(dimensions, weights);
7690 return mImpl->addIdentity(input);
7705 return mImpl->addCast(input, toType);
7720 mImpl->removeTensor(tensor);
7732 mImpl->unmarkOutput(tensor);
7751 return mImpl->addSlice(input, start, size, stride);
7775 mImpl->setName(name);
7789 return mImpl->getName();
7805 return mImpl->addShape(input);
7815 return mImpl->getFlags();
7827 return mImpl->getFlag(networkDefinitionCreationFlag);
7844 return mImpl->markOutputForShapes(tensor);
7856 return mImpl->unmarkOutputForShapes(tensor);
7874 return mImpl->addParametricReLU(input, slope);
7897 return mImpl->addConvolutionNd(input, nbOutputMaps, kernelSize, kernelWeights, biasWeights);
7916 return mImpl->addPoolingNd(input, type, windowSize);
7939 return mImpl->addDeconvolutionNd(input, nbOutputMaps, kernelSize, kernelWeights, biasWeights);
7976 return mImpl->addScaleNd(input, mode, shift, scale, power, channelAxis);
7992 return mImpl->addResize(input);
8006 return mImpl->addLoop();
8021 return mImpl->addIfConditional();
8060 return mImpl->addSelect(condition, thenInput, elseInput);
8077 return mImpl->addAssertion(condition, message);
8103 return mImpl->addFillV2(dimensions, op, outputType);
8119 return mImpl->addPaddingNd(input, prePadding, postPadding);
8143 return mImpl->setWeightsName(weights, name);
8162 mImpl->setErrorRecorder(recorder);
8177 return mImpl->getErrorRecorder();
8200 return mImpl->addDequantizeV2(input, scale, outputType);
8220 return mImpl->addScatter(data, indices, updates, mode);
8244 return mImpl->addQuantizeV2(input, scale, outputType);
8272 return mImpl->addDynamicQuantize(input, axis, blockSize, outputType, scaleType);
8287 return mImpl->addEinsum(inputs, nbInputs, equation);
8305 return mImpl->addGridSample(input, grid);
8327 return mImpl->addNMS(boxes, scores, maxOutputBoxesPerClass);
8347 return mImpl->addNMSV2(boxes, scores, maxOutputBoxesPerClass, indicesType);
8364 return mImpl->addReverseSequence(input, sequenceLens);
8390 return mImpl->addNormalization(input, scale, bias, axesMask);
8412 return mImpl->addCumulative(input, axis, operation, exclusive, reverse);
8440 return mImpl->addAttention(query, key, value, normOp, causal);
8451 return mImpl->getBuilder();
8464 return mImpl->markWeightsRefittable(name);
8476 return mImpl->unmarkWeightsRefittable(name);
8489 return mImpl->areWeightsMarkedRefittable(name);
8508 return mImpl->addSqueeze(input, axes);
8529 return mImpl->addUnsqueeze(input, axes);
8576 static constexpr int32_t kVALUE = 2;
8771#if ENABLE_FEATURE_DISABLE_RUNTIME_ALLOCATION
8778 kREQUIRE_USER_ALLOCATION = 29,
8791#if ENABLE_FEATURE_DISABLE_RUNTIME_ALLOCATION
8835 static constexpr uint64_t kINVALID_TACTIC_HASH = UINT64_MAX;
8870 return mImpl->serialize();
8894 return mImpl->combine(inputCache, ignoreMismatch);
8904 return mImpl->reset();
8921 int64_t
queryKeys(TimingCacheKey* keyBuffer, int64_t capacity)
const noexcept
8923 return mImpl->queryKeys(keyBuffer, capacity);
8938 TimingCacheValue
query(TimingCacheKey
const& key)
const noexcept
8940 return mImpl->query(key);
8960 bool update(TimingCacheKey
const& key, TimingCacheValue
const& value)
noexcept
8962 return mImpl->update(key, value);
9083 static constexpr int32_t kVALUE = 3;
9135 static constexpr int32_t kVALUE = 3;
9197 static constexpr int32_t kVALUE = 4;
9236 virtual void phaseStart(
char const* phaseName,
char const* parentPhase, int32_t nbSteps)
noexcept = 0;
9250 virtual bool stepComplete(
char const* phaseName, int32_t step)
noexcept = 0;
9309 virtual
void setAvgTimingIterations(int32_t avgTiming) noexcept
9311 mImpl->setAvgTimingIterations(avgTiming);
9323 return mImpl->getAvgTimingIterations();
9336 mImpl->setEngineCapability(capability);
9348 return mImpl->getEngineCapability();
9365 mImpl->setFlags(builderFlags);
9377 return mImpl->getFlags();
9389 mImpl->clearFlag(builderFlag);
9401 mImpl->setFlag(builderFlag);
9413 return mImpl->getFlag(builderFlag);
9430 mImpl->setDeviceType(layer, deviceType);
9440 return mImpl->getDeviceType(layer);
9452 return mImpl->isDeviceTypeSet(layer);
9462 mImpl->resetDeviceType(layer);
9472 return mImpl->canRunOnDLA(layer);
9488 mImpl->setDLACore(dlaCore);
9498 return mImpl->getDLACore();
9509 mImpl->setDefaultDeviceType(deviceType);
9519 return mImpl->getDefaultDeviceType();
9541 return mImpl->setProfileStream(stream);
9553 return mImpl->getProfileStream();
9570 return mImpl->addOptimizationProfile(profile);
9583 return mImpl->getNbOptimizationProfiles();
9595 mImpl->setProfilingVerbosity(verbosity);
9608 return mImpl->getProfilingVerbosity();
9630 return mImpl->setTacticSources(tacticSources);
9645 return mImpl->getTacticSources();
9667 return mImpl->createTimingCache(blob, size);
9692 return mImpl->setTimingCache(cache, ignoreMismatch);
9704 return mImpl->getTimingCache();
9736 mImpl->setMemoryPoolLimit(pool, poolSize);
9755 return mImpl->getMemoryPoolLimit(pool);
9773 mImpl->setPreviewFeature(feature, enable);
9787 return mImpl->getPreviewFeature(feature);
9820 mImpl->setBuilderOptimizationLevel(level);
9832 return mImpl->getBuilderOptimizationLevel();
9849 mImpl->setHardwareCompatibilityLevel(hardwareCompatibilityLevel);
9862 return mImpl->getHardwareCompatibilityLevel();
9875 mImpl->setPluginsToSerialize(paths, nbPaths);
9888 return mImpl->getPluginToSerialize(index);
9898 return mImpl->getNbPluginsToSerialize();
9927 mImpl->setMaxAuxStreams(nbStreams);
9937 return mImpl->getMaxAuxStreams();
9953 return mImpl->setProgressMonitor(monitor);
9963 return mImpl->getProgressMonitor();
9979 mImpl->setRuntimePlatform(runtimePlatform);
9991 return mImpl->getRuntimePlatform();
10003 mImpl->setMaxNbTactics(maxNbTactics);
10015 return mImpl->getMaxNbTactics();
10031 return mImpl->setTilingOptimizationLevel(level);
10043 return mImpl->getTilingOptimizationLevel();
10059 return mImpl->setL2LimitForTiling(size);
10071 return mImpl->getL2LimitForTiling();
10090 return mImpl->setNbComputeCapabilities(maxNbComputeCapabilities);
10102 return mImpl->getNbComputeCapabilities();
10120 return mImpl->setComputeCapability(computeCapability, index);
10134 return mImpl->getComputeCapability(index);
10202 int32_t getMaxDLABatchSize() const noexcept
10204 return mImpl->getMaxDLABatchSize();
10212 return mImpl->getNbDLACores();
10230 mImpl->setGpuAllocator(allocator);
10244 return mImpl->createBuilderConfig();
10270 return mImpl->createNetworkV2(flags);
10285 return mImpl->createOptimizationProfile();
10304 mImpl->setErrorRecorder(recorder);
10319 return mImpl->getErrorRecorder();
10346 return mImpl->buildSerializedNetwork(network, config);
10368 return mImpl->buildSerializedNetworkToStream(network, config, writer);
10391 return mImpl->isNetworkSupported(network, config);
10401 return mImpl->getLogger();
10417 return mImpl->setMaxThreads(maxThreads);
10431 return mImpl->getMaxThreads();
10441 return mImpl->getPluginRegistry();
10454extern "C" TENSORRTAPI void* createInferBuilder_INTERNAL(
void* logger, int32_t version)
noexcept;
#define TENSORRTAPI
Definition: NvInferRuntimeBase.h:69
#define NV_TENSORRT_VERSION
Definition: NvInferRuntimeBase.h:101
#define TRT_DEPRECATED
Definition: NvInferRuntimeBase.h:42
#define TRT_DEPRECATED_ENUM
Definition: NvInferRuntimeBase.h:43
Definition: NvInferRuntimeBase.h:218
static constexpr int32_t MAX_DIMS
The maximum rank (number of dimensions) supported for a tensor.
Definition: NvInferRuntimeBase.h:221
An Activation layer in a network definition.
Definition: NvInfer.h:1265
void setBeta(float beta) noexcept
Set the beta parameter (must be finite).
Definition: NvInfer.h:1313
void setActivationType(ActivationType type) noexcept
Set the type of activation to be performed.
Definition: NvInfer.h:1274
ActivationType getActivationType() const noexcept
Get the type of activation to be performed.
Definition: NvInfer.h:1284
float getAlpha() const noexcept
Get the alpha parameter.
Definition: NvInfer.h:1322
virtual ~IActivationLayer() noexcept=default
float getBeta() const noexcept
Get the beta parameter.
Definition: NvInfer.h:1331
void setAlpha(float alpha) noexcept
Set the alpha parameter (must be finite).
Definition: NvInfer.h:1299
An assertion layer in a network.
Definition: NvInfer.h:4934
void setMessage(char const *message) noexcept
Set the message to print if the assertion fails.
Definition: NvInfer.h:4944
char const * getMessage() const noexcept
Return the assertion message.
Definition: NvInfer.h:4954
virtual ~IAssertionLayer() noexcept=default
This is a base class for Attention boundary layers.
Definition: NvInfer.h:6670
IAttention * getAttention() const noexcept
Get a pointer to the IAttention associated with this boundary layer.
Definition: NvInfer.h:6675
virtual ~IAttentionBoundaryLayer() noexcept=default
Helper for constructing an attention that consumes query, key and value tensors.
Definition: NvInfer.h:6784
ITensor * getMask() noexcept
Get the optional mask in attention.
Definition: NvInfer.h:6834
bool setDecomposable(bool decomposable) noexcept
Set whether the attention can be decomposed to use multiple kernels if no fused kernel support found.
Definition: NvInfer.h:6871
bool setName(char const *name) noexcept
Set the name of the attention.
Definition: NvInfer.h:6961
bool getDecomposable() const noexcept
Get whether the attention can be decomposed to use multiple kernels if no fused kernel support found.
Definition: NvInfer.h:6884
ITensor * getInput(int32_t index) const noexcept
Get the IAttention input corresponding to the given index.
Definition: NvInfer.h:6924
ITensor * getOutput(int32_t index) const noexcept
Get the IAttention output corresponding to the given index. IAttention has only one output.
Definition: NvInfer.h:6944
int32_t getNbOutputs() const noexcept
Get the number of outputs of a layer. IAttention has one output.
Definition: NvInfer.h:6932
int32_t getNbInputs() const noexcept
Get the number of inputs of IAttention. IAttention has three inputs.
Definition: NvInfer.h:6912
bool setCausal(bool isCausal) noexcept
Set whether the attention will run a causal inference. Cannot be used together with setMask().
Definition: NvInfer.h:6847
bool setNormalizationOperation(AttentionNormalizationOp op) noexcept
Set the normalization operation for the attention.
Definition: NvInfer.h:6793
char const * getName() const noexcept
Return the name of the attention.
Definition: NvInfer.h:6973
bool setNormalizationQuantizeToType(DataType type) noexcept
Set the datatype the attention normalization is quantized to.
Definition: NvInfer.h:7013
AttentionNormalizationOp getNormalizationOperation() const noexcept
Get the normalization operation for the attention.
Definition: NvInfer.h:6805
bool setNormalizationQuantizeScale(ITensor &tensor) noexcept
Set the quantization scale for the attention normalization output.
Definition: NvInfer.h:6989
DataType getNormalizationQuantizeToType() const noexcept
Get the datatype the attention normalization is quantized to.
Definition: NvInfer.h:7025
ITensor * getNormalizationQuantizeScale() const noexcept
Get the quantization scale for the attention normalization output.
Definition: NvInfer.h:7000
bool setInput(int32_t index, ITensor &input) noexcept
Append or replace an input of this layer with a specific tensor.
Definition: NvInfer.h:6903
bool setMask(ITensor &mask) noexcept
Set whether a mask will be used for the normalization operation.
Definition: NvInfer.h:6822
bool getCausal() const noexcept
Get whether the attention will run a causal inference.
Definition: NvInfer.h:6859
apiv::VAttention * mImpl
Definition: NvInfer.h:7032
virtual ~IAttention() noexcept=default
This layer represents an output of an IAttention.
Definition: NvInfer.h:6731
virtual ~IAttentionOutputLayer() noexcept=default
Holds properties for configuring a builder to produce an engine.
Definition: NvInfer.h:9297
void setMemoryPoolLimit(MemoryPoolType pool, std::size_t poolSize) noexcept
Set the memory size for the memory pool.
Definition: NvInfer.h:9734
bool setComputeCapability(ComputeCapability computeCapability, int32_t index) noexcept
Set one compute capability for runtime execution.
Definition: NvInfer.h:10118
bool setNbComputeCapabilities(int32_t maxNbComputeCapabilities) noexcept
Set the number of compute capabilities.
Definition: NvInfer.h:10088
TRT_DEPRECATED bool setTimingCache(ITimingCache const &cache, bool ignoreMismatch) noexcept
Attach a timing cache to IBuilderConfig.
Definition: NvInfer.h:9690
void setPreviewFeature(PreviewFeature feature, bool enable) noexcept
Enable or disable a specific preview feature.
Definition: NvInfer.h:9771
bool getPreviewFeature(PreviewFeature feature) const noexcept
Get status of preview feature.
Definition: NvInfer.h:9785
int32_t getBuilderOptimizationLevel() noexcept
Get builder optimization level.
Definition: NvInfer.h:9830
bool setTacticSources(TacticSources tacticSources) noexcept
Set tactic sources.
Definition: NvInfer.h:9628
void setPluginsToSerialize(char const *const *paths, int32_t nbPaths) noexcept
Set the plugin libraries to be serialized with version-compatible engines.
Definition: NvInfer.h:9873
bool setTilingOptimizationLevel(TilingOptimizationLevel level) noexcept
Set the Tiling optimization level.
Definition: NvInfer.h:10029
bool setL2LimitForTiling(int64_t size) noexcept
Set the L2 cache usage limit for Tiling optimization.
Definition: NvInfer.h:10057
std::size_t getMemoryPoolLimit(MemoryPoolType pool) const noexcept
Get the memory size limit of the memory pool.
Definition: NvInfer.h:9753
int32_t getDLACore() const noexcept
Get the DLA core that the engine executes on.
Definition: NvInfer.h:9496
int32_t getNbPluginsToSerialize() const noexcept
Get the number of plugin library paths to be serialized with version-compatible engines.
Definition: NvInfer.h:9896
void setDeviceType(ILayer const *layer, DeviceType deviceType) noexcept
Set the device that this layer must execute on.
Definition: NvInfer.h:9428
void setEngineCapability(EngineCapability capability) noexcept
Configure the builder to target specified EngineCapability flow.
Definition: NvInfer.h:9334
int32_t getMaxAuxStreams() const noexcept
Get the maximum number of auxiliary streams that TRT is allowed to use.
Definition: NvInfer.h:9935
bool getFlag(BuilderFlag builderFlag) const noexcept
Returns true if the build mode flag is set.
Definition: NvInfer.h:9411
void setMaxNbTactics(int32_t maxNbTactics) noexcept
Set the maximum number of tactics to time when there is a choice of tactics.
Definition: NvInfer.h:10001
int64_t getL2LimitForTiling() const noexcept
Get the L2 cache usage limit for tiling optimization.
Definition: NvInfer.h:10069
void setProgressMonitor(IProgressMonitor *monitor) noexcept
Sets the progress monitor for building a network.
Definition: NvInfer.h:9951
void setProfilingVerbosity(ProfilingVerbosity verbosity) noexcept
Set verbosity level of layer information exposed in NVTX annotations and IEngineInspector.
Definition: NvInfer.h:9593
int32_t getNbOptimizationProfiles() const noexcept
Get number of optimization profiles.
Definition: NvInfer.h:9581
void reset() noexcept
Resets the builder configuration to defaults.
Definition: NvInfer.h:9527
char const * getPluginToSerialize(int32_t index) const noexcept
Get the plugin library path to be serialized with version-compatible engines.
Definition: NvInfer.h:9886
EngineCapability getEngineCapability() const noexcept
Query EngineCapability flow configured for the builder.
Definition: NvInfer.h:9346
RuntimePlatform getRuntimePlatform() const noexcept
Get the target platform for runtime execution.
Definition: NvInfer.h:9989
DeviceType getDefaultDeviceType() const noexcept
Get the default DeviceType which was set by setDefaultDeviceType.
Definition: NvInfer.h:9517
void setRuntimePlatform(RuntimePlatform runtimePlatform) noexcept
Set the target platform for runtime execution.
Definition: NvInfer.h:9977
int32_t getMaxNbTactics() const noexcept
Query the maximum number of tactics timed when there is a choice.
Definition: NvInfer.h:10013
BuilderFlags getFlags() const noexcept
Get the build mode flags for this builder config. Defaults to 0.
Definition: NvInfer.h:9375
void setFlags(BuilderFlags builderFlags) noexcept
Set the build mode flags to turn on builder options for this network.
Definition: NvInfer.h:9363
TacticSources getTacticSources() const noexcept
Get tactic sources.
Definition: NvInfer.h:9643
void resetDeviceType(ILayer const *layer) noexcept
reset the DeviceType for this layer
Definition: NvInfer.h:9460
ComputeCapability getComputeCapability(int32_t index) const noexcept
Get one compute capability for runtime execution.
Definition: NvInfer.h:10132
void setDLACore(int32_t dlaCore) noexcept
Sets the DLA core used by the network. Defaults to -1.
Definition: NvInfer.h:9486
HardwareCompatibilityLevel getHardwareCompatibilityLevel() const noexcept
Get the hardware compatibility level.
Definition: NvInfer.h:9860
int32_t getNbComputeCapabilities() const noexcept
Get the number of compute capabilities.
Definition: NvInfer.h:10100
void clearFlag(BuilderFlag builderFlag) noexcept
clear a single build mode flag.
Definition: NvInfer.h:9387
int32_t addOptimizationProfile(IOptimizationProfile const *profile) noexcept
Add an optimization profile.
Definition: NvInfer.h:9568
IProgressMonitor * getProgressMonitor() const noexcept
Definition: NvInfer.h:9961
apiv::VBuilderConfig * mImpl
Definition: NvInfer.h:10138
int32_t getAvgTimingIterations() const noexcept
Query the number of averaging iterations.
Definition: NvInfer.h:9321
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:9507
void setFlag(BuilderFlag builderFlag) noexcept
Set a single build mode flag.
Definition: NvInfer.h:9399
TRT_DEPRECATED nvinfer1::ITimingCache * createTimingCache(void const *blob, std::size_t size) const noexcept
Create timing cache.
Definition: NvInfer.h:9665
virtual ~IBuilderConfig() noexcept=default
DeviceType getDeviceType(ILayer const *layer) const noexcept
Get the device that this layer executes on.
Definition: NvInfer.h:9438
bool canRunOnDLA(ILayer const *layer) const noexcept
Checks if a layer can run on DLA.
Definition: NvInfer.h:9470
cudaStream_t getProfileStream() const noexcept
Get the CUDA stream that is used to profile this network.
Definition: NvInfer.h:9551
void setHardwareCompatibilityLevel(HardwareCompatibilityLevel hardwareCompatibilityLevel) noexcept
Set the hardware compatibility level.
Definition: NvInfer.h:9847
TilingOptimizationLevel getTilingOptimizationLevel() const noexcept
Get the Tiling optimization level.
Definition: NvInfer.h:10041
void setMaxAuxStreams(int32_t nbStreams) noexcept
Set the maximum number of auxiliary streams that TRT is allowed to use.
Definition: NvInfer.h:9925
ProfilingVerbosity getProfilingVerbosity() const noexcept
Get verbosity level of layer information exposed in NVTX annotations and IEngineInspector.
Definition: NvInfer.h:9606
TRT_DEPRECATED nvinfer1::ITimingCache const * getTimingCache() const noexcept
Get the pointer to the timing cache from current IBuilderConfig.
Definition: NvInfer.h:9702
bool isDeviceTypeSet(ILayer const *layer) const noexcept
whether the DeviceType has been explicitly set for this layer
Definition: NvInfer.h:9450
void setBuilderOptimizationLevel(int32_t level) noexcept
Set builder optimization level.
Definition: NvInfer.h:9818
void setProfileStream(const cudaStream_t stream) noexcept
Set the CUDA stream that is used to profile this network.
Definition: NvInfer.h:9539
Builds an engine from a network definition.
Definition: NvInfer.h:10191
int32_t getNbDLACores() const noexcept
Return the number of DLA engines available to this builder.
Definition: NvInfer.h:10210
IErrorRecorder * getErrorRecorder() const noexcept
get the ErrorRecorder assigned to this interface.
Definition: NvInfer.h:10317
apiv::VBuilder * mImpl
Definition: NvInfer.h:10445
ILogger * getLogger() const noexcept
get the logger with which the builder was created
Definition: NvInfer.h:10399
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:10389
int32_t getMaxThreads() const noexcept
get the maximum number of threads that can be used by the builder.
Definition: NvInfer.h:10429
IPluginRegistry & getPluginRegistry() noexcept
get the local plugin registry that can be used by the builder.
Definition: NvInfer.h:10439
nvinfer1::IOptimizationProfile * createOptimizationProfile() noexcept
Create a new optimization profile.
Definition: NvInfer.h:10283
void setGpuAllocator(IGpuAllocator *allocator) noexcept
Set the GPU allocator.
Definition: NvInfer.h:10228
nvinfer1::INetworkDefinition * createNetworkV2(NetworkDefinitionCreationFlags flags) noexcept
Create a network definition object.
Definition: NvInfer.h:10268
nvinfer1::IBuilderConfig * createBuilderConfig() noexcept
Create a builder configuration object.
Definition: NvInfer.h:10242
void reset() noexcept
Resets the builder state to default values.
Definition: NvInfer.h:10325
bool setMaxThreads(int32_t maxThreads) noexcept
Set the maximum number of threads.
Definition: NvInfer.h:10415
void setErrorRecorder(IErrorRecorder *recorder) noexcept
Set the ErrorRecorder for this interface.
Definition: NvInfer.h:10302
nvinfer1::IHostMemory * buildSerializedNetwork(INetworkDefinition &network, IBuilderConfig &config) noexcept
Builds and serializes a network for the given INetworkDefinition and IBuilderConfig.
Definition: NvInfer.h:10344
virtual ~IBuilder() noexcept=default
bool buildSerializedNetworkToStream(INetworkDefinition &network, IBuilderConfig &config, IStreamWriter &writer) noexcept
Builds and serializes a network into stream for the given INetworkDefinition and IBuilderConfig.
Definition: NvInfer.h:10365
A cast layer in a network.
Definition: NvInfer.h:3795
virtual ~ICastLayer() noexcept=default
apiv::VCastLayer * mImpl
Definition: NvInfer.h:3821
DataType getToType() const noexcept
Return cast layer output type.
Definition: NvInfer.h:3815
void setToType(DataType toType) noexcept
Set cast layer output type.
Definition: NvInfer.h:3804
A concatenation layer in a network definition.
Definition: NvInfer.h:1975
void setAxis(int32_t axis) noexcept
Set the axis along which concatenation occurs.
Definition: NvInfer.h:1988
int32_t getAxis() const noexcept
Get the axis along which concatenation occurs.
Definition: NvInfer.h:1998
virtual ~IConcatenationLayer() noexcept=default
This layer represents a condition input to an IIfConditional.
Definition: NvInfer.h:4458
virtual ~IConditionLayer() noexcept=default
Layer that represents a constant value.
Definition: NvInfer.h:3834
void setWeights(Weights weights) noexcept
Set the weights for the layer.
Definition: NvInfer.h:3844
Weights getWeights() const noexcept
Get the weights for the layer.
Definition: NvInfer.h:3854
void setDimensions(Dims const &dimensions) noexcept
Set the dimensions for the layer.
Definition: NvInfer.h:3866
apiv::VConstantLayer * mImpl
Definition: NvInfer.h:3884
virtual ~IConstantLayer() noexcept=default
Dims getDimensions() const noexcept
Get the dimensions for the layer.
Definition: NvInfer.h:3878
A convolution layer in a network definition.
Definition: NvInfer.h:945
Dims getPrePadding() const noexcept
Get the pre-padding.
Definition: NvInfer.h:1070
Weights getBiasWeights() const noexcept
Get the bias weights for the convolution.
Definition: NvInfer.h:1043
void setPaddingMode(PaddingMode paddingMode) noexcept
Set the padding mode.
Definition: NvInfer.h:1111
void setDilationNd(Dims const &dilation) noexcept
Set the multi-dimension dilation of the convolution.
Definition: NvInfer.h:1215
Dims getPaddingNd() const noexcept
Get the multi-dimension padding of the convolution.
Definition: NvInfer.h:1201
Dims getStrideNd() const noexcept
Get the multi-dimension stride of the convolution.
Definition: NvInfer.h:1171
Weights getKernelWeights() const noexcept
Get the kernel weights of the convolution.
Definition: NvInfer.h:1018
void setStrideNd(Dims const &stride) noexcept
Set the multi-dimension stride of the convolution.
Definition: NvInfer.h:1161
Dims getDilationNd() const noexcept
Get the multi-dimension dilation of the convolution.
Definition: NvInfer.h:1225
int64_t getNbOutputMaps() const noexcept
Get the number of output maps for the convolution.
Definition: NvInfer.h:964
void setKernelWeights(Weights weights) noexcept
Set the kernel weights for the convolution.
Definition: NvInfer.h:1008
Dims getPostPadding() const noexcept
Get the post-padding.
Definition: NvInfer.h:1097
int64_t getNbGroups() const noexcept
Get the number of groups of the convolution.
Definition: NvInfer.h:994
PaddingMode getPaddingMode() const noexcept
Get the padding mode.
Definition: NvInfer.h:1123
virtual ~IConvolutionLayer() noexcept=default
void setNbGroups(int64_t nbGroups) noexcept
Set the number of groups for a convolution.
Definition: NvInfer.h:984
void setNbOutputMaps(int64_t nbOutputMaps) noexcept
Set the number of output maps for the convolution.
Definition: NvInfer.h:954
void setBiasWeights(Weights weights) noexcept
Set the bias weights for the convolution.
Definition: NvInfer.h:1033
Dims getKernelSizeNd() const noexcept
Get the multi-dimension kernel size of the convolution.
Definition: NvInfer.h:1146
void setPaddingNd(Dims const &padding) noexcept
Set the multi-dimension padding of the convolution.
Definition: NvInfer.h:1189
void setPrePadding(Dims const &padding) noexcept
Set the multi-dimension pre-padding of the convolution.
Definition: NvInfer.h:1060
void setPostPadding(Dims const &padding) noexcept
Set the multi-dimension post-padding of the convolution.
Definition: NvInfer.h:1087
void setKernelSizeNd(Dims const &kernelSize) noexcept
Set the multi-dimension kernel size of the convolution.
Definition: NvInfer.h:1136
Layer that represents a cumulative operation across a tensor.
Definition: NvInfer.h:6552
bool setOperation(CumulativeOperation op) noexcept
Set the cumulative operation for the layer.
Definition: NvInfer.h:6563
void setReverse(bool reverse) noexcept
Specify whether the cumulative operation should be applied backward.
Definition: NvInfer.h:6611
apiv::VCumulativeLayer * mImpl
Definition: NvInfer.h:6629
bool getExclusive() const noexcept
Get whether it is exclusive accumulation or inclusive accumulation.
Definition: NvInfer.h:6599
virtual ~ICumulativeLayer() noexcept=default
bool getReverse() const noexcept
Get the boolean that specifies whether the cumulative operation should be applied backward.
Definition: NvInfer.h:6623
void setExclusive(bool exclusive) noexcept
Set whether it is an exclusive accumulation or inclusive accumulation.
Definition: NvInfer.h:6587
CumulativeOperation getOperation() const noexcept
Get the cumulative operation for the layer.
Definition: NvInfer.h:6575
A deconvolution layer in a network definition.
Definition: NvInfer.h:2016
void setBiasWeights(Weights weights) noexcept
Set the bias weights for the deconvolution.
Definition: NvInfer.h:2104
int64_t getNbGroups() const noexcept
Get the number of groups for a deconvolution.
Definition: NvInfer.h:2065
Weights getKernelWeights() const noexcept
Get the kernel weights for the deconvolution.
Definition: NvInfer.h:2089
void setPrePadding(Dims const &padding) noexcept
Set the multi-dimension pre-padding of the deconvolution.
Definition: NvInfer.h:2131
Dims getStrideNd() const noexcept
Get the multi-dimension stride of the deconvolution.
Definition: NvInfer.h:2246
Dims getDilationNd() const noexcept
Get the multi-dimension dilation of the deconvolution.
Definition: NvInfer.h:2312
Weights getBiasWeights() const noexcept
Get the bias weights for the deconvolution.
Definition: NvInfer.h:2114
void setKernelWeights(Weights weights) noexcept
Set the kernel weights for the deconvolution.
Definition: NvInfer.h:2079
int64_t getNbOutputMaps() const noexcept
Get the number of output feature maps for the deconvolution.
Definition: NvInfer.h:2035
void setStrideNd(Dims const &stride) noexcept
Set the multi-dimension stride of the deconvolution.
Definition: NvInfer.h:2236
Dims getPostPadding() const noexcept
Get the padding.
Definition: NvInfer.h:2168
Dims getKernelSizeNd() const noexcept
Get the multi-dimension kernel size of the deconvolution.
Definition: NvInfer.h:2219
void setPostPadding(Dims const &padding) noexcept
Set the multi-dimension post-padding of the deconvolution.
Definition: NvInfer.h:2158
void setKernelSizeNd(Dims const &kernelSize) noexcept
Set the multi-dimension kernel size of the deconvolution.
Definition: NvInfer.h:2209
virtual ~IDeconvolutionLayer() noexcept=default
void setPaddingNd(Dims const &padding) noexcept
Set the multi-dimension padding of the deconvolution.
Definition: NvInfer.h:2264
void setNbOutputMaps(int64_t nbOutputMaps) noexcept
Set the number of output feature maps for the deconvolution.
Definition: NvInfer.h:2025
Dims getPaddingNd() const noexcept
Get the multi-dimension padding of the deconvolution.
Definition: NvInfer.h:2276
void setDilationNd(Dims const &dilation) noexcept
Set the multi-dimension dilation of the deconvolution.
Definition: NvInfer.h:2302
void setPaddingMode(PaddingMode paddingMode) noexcept
Set the padding mode.
Definition: NvInfer.h:2182
void setNbGroups(int64_t nbGroups) noexcept
Set the number of groups for a deconvolution.
Definition: NvInfer.h:2055
Dims getPrePadding() const noexcept
Get the pre-padding.
Definition: NvInfer.h:2141
PaddingMode getPaddingMode() const noexcept
Get the padding mode.
Definition: NvInfer.h:2194
A Dequantize layer in a network definition.
Definition: NvInfer.h:5519
void setToType(DataType toType) noexcept
Set the Dequantize layer output type.
Definition: NvInfer.h:5556
virtual ~IDequantizeLayer() noexcept=default
int32_t getAxis() const noexcept
Get the quantization axis.
Definition: NvInfer.h:5529
DataType getToType() const noexcept
Return the Dequantize layer output type.
Definition: NvInfer.h:5568
void setAxis(int32_t axis) noexcept
Set the quantization axis.
Definition: NvInfer.h:5540
A network layer to perform dynamic quantization.
Definition: NvInfer.h:5596
int32_t getAxis() const noexcept
Get the axis along which blocking occurs.
Definition: NvInfer.h:5685
int32_t getBlockSize() const noexcept
Get the size of the quantization block.
Definition: NvInfer.h:5708
DataType getScaleType() const noexcept
Return the scale factors data type.
Definition: NvInfer.h:5662
void setScaleType(DataType scaleType) noexcept
Set the data type of the scale factors used to quantize the data.
Definition: NvInfer.h:5649
DataType getToType() const noexcept
Return DynamicQuantizeLayer's quantized output type.
Definition: NvInfer.h:5636
virtual ~IDynamicQuantizeLayer() noexcept=default
void setToType(DataType toType) noexcept
Set DynamicQuantizeLayer's quantized output type.
Definition: NvInfer.h:5623
void setAxis(int32_t axis) noexcept
Set the axis along which block quantization occurs.
Definition: NvInfer.h:5675
void setBlockSize(int32_t size) noexcept
Set the size of the quantization block.
Definition: NvInfer.h:5698
An Einsum layer in a network.
Definition: NvInfer.h:5753
bool setEquation(char const *equation) noexcept
Set the equation. The equation is a comma-separated list of subscript labels, where each label refers...
Definition: NvInfer.h:5764
virtual ~IEinsumLayer() noexcept=default
char const * getEquation() const noexcept
Return the equation.
Definition: NvInfer.h:5774
A elementwise layer in a network definition.
Definition: NvInfer.h:2386
virtual ~IElementWiseLayer() noexcept=default
apiv::VElementWiseLayer * mImpl
Definition: NvInfer.h:2415
ElementWiseOperation getOperation() const noexcept
Get the binary operation for the layer.
Definition: NvInfer.h:2409
void setOperation(ElementWiseOperation op) noexcept
Set the binary operation for the layer.
Definition: NvInfer.h:2397
Generate a tensor according to a specified mode.
Definition: NvInfer.h:5045
bool isAlphaBetaInt64() const noexcept
Return true if alpha/beta have type int64, false if they have type double.
Definition: NvInfer.h:5277
FillOperation getOperation() const noexcept
Get the fill operation for the layer.
Definition: NvInfer.h:5091
void setOperation(FillOperation op) noexcept
Set the fill operation for the layer.
Definition: NvInfer.h:5081
DataType getToType() const noexcept
Get the fill layer output type.
Definition: NvInfer.h:5306
void setAlphaInt64(int64_t alpha) noexcept
Set the alpha parameter with int64 datatype.
Definition: NvInfer.h:5220
void setBetaInt64(int64_t beta) noexcept
Set the beta parameter with int64 datatype.
Definition: NvInfer.h:5254
void setBeta(double beta) noexcept
Set the beta parameter.
Definition: NvInfer.h:5144
int64_t getAlphaInt64() const noexcept
Get the value of alpha parameter with int64 datatype.
Definition: NvInfer.h:5235
int64_t getBetaInt64() const noexcept
Get the value of beta parameter with int64 datatype.
Definition: NvInfer.h:5269
double getAlpha() const noexcept
Get the value of alpha parameter.
Definition: NvInfer.h:5125
void setDimensions(Dims const &dimensions) noexcept
Set the output tensor's dimensions.
Definition: NvInfer.h:5056
void setAlpha(double alpha) noexcept
Set the alpha parameter.
Definition: NvInfer.h:5110
void setToType(DataType toType) noexcept
Set the fill layer output type.
Definition: NvInfer.h:5294
Dims getDimensions() const noexcept
Get the output tensor's dimensions.
Definition: NvInfer.h:5071
double getBeta() const noexcept
Get the value of beta parameter.
Definition: NvInfer.h:5159
virtual ~IFillLayer() noexcept=default
A Gather layer in a network definition. Supports several kinds of gathering.
Definition: NvInfer.h:2519
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:2530
void setNbElementWiseDims(int32_t elementWiseDims) noexcept
Set the number of leading dimensions of indices tensor to be handled elementwise.
Definition: NvInfer.h:2565
apiv::VGatherLayer * mImpl
Definition: NvInfer.h:2601
int32_t getNbElementWiseDims() const noexcept
Get the number of leading dimensions of indices tensor to be handled elementwise.
Definition: NvInfer.h:2575
void setMode(GatherMode mode) noexcept
Set the gather mode.
Definition: NvInfer.h:2585
int32_t getGatherAxis() const noexcept
Get the axis to gather on.
Definition: NvInfer.h:2542
GatherMode getMode() const noexcept
Get the gather mode.
Definition: NvInfer.h:2595
virtual ~IGatherLayer() noexcept=default
A GridSample layer in a network definition.
Definition: NvInfer.h:5975
void setInterpolationMode(InterpolationMode mode) noexcept
Set the grid sample interpolation mode.
Definition: NvInfer.h:5982
bool setSampleMode(SampleMode mode) noexcept
Set the sample mode.
Definition: NvInfer.h:6028
void setAlignCorners(bool alignCorners) noexcept
Set the align corners mode.
Definition: NvInfer.h:6004
apiv::VGridSampleLayer * mImpl
Definition: NvInfer.h:6046
SampleMode getSampleMode() const noexcept
Get the sample mode.
Definition: NvInfer.h:6040
InterpolationMode getInterpolationMode() const noexcept
Get the grid sample interpolation mode.
Definition: NvInfer.h:5994
bool getAlignCorners() const noexcept
Get the align corners mode.
Definition: NvInfer.h:6016
virtual ~IGridSampleLayer() noexcept=default
Class to handle library allocated memory that is accessible to the user.
Definition: NvInferRuntime.h:142
A layer that represents the identity function.
Definition: NvInfer.h:3782
apiv::VIdentityLayer * mImpl
Definition: NvInfer.h:3784
virtual ~IIdentityLayer() noexcept=default
This is a base class for Conditional boundary layers.
Definition: NvInfer.h:4437
IIfConditional * getConditional() const noexcept
Get a pointer to the IIfConditional associated with this boundary layer.
Definition: NvInfer.h:4442
virtual ~IIfConditionalBoundaryLayer() noexcept=default
Helper for constructing conditionally-executed subgraphs.
Definition: NvInfer.h:4520
IIfConditionalInputLayer * addInput(ITensor &input) noexcept
Add an If-conditional input.
Definition: NvInfer.h:4561
char const * getName() const noexcept
Return the name of the conditional.
Definition: NvInfer.h:4586
virtual ~IIfConditional() noexcept=default
IConditionLayer * setCondition(ITensor &condition) noexcept
Set the condition tensor for this If-Conditional construct.
Definition: NvInfer.h:4531
IIfConditionalOutputLayer * addOutput(ITensor &trueSubgraphOutput, ITensor &falseSubgraphOutput) noexcept
Add an If-conditional output.
Definition: NvInfer.h:4549
void setName(char const *name) noexcept
Set the name of the conditional.
Definition: NvInfer.h:4576
This layer represents an output of an IIfConditional.
Definition: NvInfer.h:4475
virtual ~IIfConditionalOutputLayer() noexcept=default
A layer to do iterations.
Definition: NvInfer.h:4751
virtual ~IIteratorLayer() noexcept=default
void setReverse(bool reverse) noexcept
Set iteration order to be reverse.
Definition: NvInfer.h:4778
bool getReverse() const noexcept
Check if the iteration order is reverse.
Definition: NvInfer.h:4788
int32_t getAxis() const noexcept
Get axis being iterated over.
Definition: NvInfer.h:4764
void setAxis(int32_t axis) noexcept
Set axis to iterate over.
Definition: NvInfer.h:4756
A LRN layer in a network definition.
Definition: NvInfer.h:1630
int64_t getWindowSize() const noexcept
Get the LRN window size.
Definition: NvInfer.h:1651
float getAlpha() const noexcept
Get the LRN alpha value.
Definition: NvInfer.h:1673
void setWindowSize(int64_t windowSize) noexcept
Set the LRN window size.
Definition: NvInfer.h:1641
void setK(float k) noexcept
Set the LRN K value.
Definition: NvInfer.h:1707
void setAlpha(float alpha) noexcept
Set the LRN alpha value.
Definition: NvInfer.h:1663
void setBeta(float beta) noexcept
Set the LRN beta value.
Definition: NvInfer.h:1685
virtual ~ILRNLayer() noexcept=default
float getBeta() const noexcept
Get the LRN beta value.
Definition: NvInfer.h:1695
float getK() const noexcept
Get the LRN K value.
Definition: NvInfer.h:1717
Base class for all layer classes in a network definition.
Definition: NvInfer.h:460
TRT_DEPRECATED void setPrecision(DataType dataType) noexcept
Set the preferred or required computational precision of this layer in a weakly-typed network.
Definition: NvInfer.h:580
TRT_DEPRECATED void setOutputType(int32_t index, DataType dataType) noexcept
Set the output type of this layer in a weakly-typed network.
Definition: NvInfer.h:668
TRT_DEPRECATED bool precisionIsSet() const noexcept
whether the computational precision has been set for this layer
Definition: NvInfer.h:606
void setMetadata(char const *metadata) noexcept
Set the metadata for this layer.
Definition: NvInfer.h:731
TRT_DEPRECATED void resetOutputType(int32_t index) noexcept
reset the output type for this layer
Definition: NvInfer.h:713
void setName(char const *name) noexcept
Set the name of a layer.
Definition: NvInfer.h:481
int32_t getNbInputs() const noexcept
Get the number of inputs of a layer.
Definition: NvInfer.h:499
char const * getMetadata() const noexcept
Get the metadata of the layer.
Definition: NvInfer.h:744
DataType getOutputType(int32_t index) const noexcept
get the output type of this layer
Definition: NvInfer.h:683
DataType getPrecision() const noexcept
get the computational precision of this layer
Definition: NvInfer.h:592
TRT_DEPRECATED bool outputTypeIsSet(int32_t index) const noexcept
whether the output type has been set for this layer
Definition: NvInfer.h:699
char const * getName() const noexcept
Return the name of a layer.
Definition: NvInfer.h:491
int32_t getNbOutputs() const noexcept
Get the number of outputs of a layer.
Definition: NvInfer.h:520
ITensor * getOutput(int32_t index) const noexcept
Get the layer output corresponding to the given index.
Definition: NvInfer.h:530
void setInput(int32_t index, ITensor &tensor) noexcept
Replace an input of this layer with a specific tensor.
Definition: NvInfer.h:547
ITensor * getInput(int32_t index) const noexcept
Get the layer input corresponding to the given index.
Definition: NvInfer.h:512
LayerType getType() const noexcept
Return the type of a layer.
Definition: NvInfer.h:467
TRT_DEPRECATED void resetPrecision() noexcept
reset the computational precision for this layer
Definition: NvInfer.h:618
virtual ~ILayer() noexcept=default
Application-implemented logging interface for the builder, refitter and runtime.
Definition: NvInferRuntime.h:1588
This is a base class for Loop boundary layers.
Definition: NvInfer.h:4414
virtual ~ILoopBoundaryLayer() noexcept=default
ILoop * getLoop() const noexcept
Get a pointer to ILoop associated with this boundary layer.
Definition: NvInfer.h:4419
Helper for creating a recurrent subgraph.
Definition: NvInfer.h:4809
void setName(char const *name) noexcept
Set the name of the loop.
Definition: NvInfer.h:4879
ITripLimitLayer * addTripLimit(ITensor &tensor, TripLimit limit) noexcept
Add a trip-count limiter, based on the given tensor.
Definition: NvInfer.h:4838
IIteratorLayer * addIterator(ITensor &tensor, int32_t axis=0, bool reverse=false) noexcept
Return layer that subscripts tensor by loop iteration.
Definition: NvInfer.h:4851
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:4864
virtual ~ILoop() noexcept=default
char const * getName() const noexcept
Return the name of the loop.
Definition: NvInfer.h:4889
IRecurrenceLayer * addRecurrence(ITensor &initialValue) noexcept
Create a recurrence layer for this loop with initialValue as its first input.
Definition: NvInfer.h:4817
An ILoopOutputLayer is the sole way to get output from a loop.
Definition: NvInfer.h:4651
virtual ~ILoopOutputLayer() noexcept=default
int32_t getAxis() const noexcept
Get axis being concatenated over.
Definition: NvInfer.h:4681
LoopOutput getLoopOutput() const noexcept
Get which kind a loop output has.
Definition: NvInfer.h:4656
void setAxis(int32_t axis) noexcept
Set where to insert the contenation axis. Ignored if getLoopOutput() is kLAST_VALUE.
Definition: NvInfer.h:4673
Layer that represents a Matrix Multiplication.
Definition: NvInfer.h:3629
apiv::VMatrixMultiplyLayer * mImpl
Definition: NvInfer.h:3657
virtual ~IMatrixMultiplyLayer() noexcept=default
MatrixOperation getOperation(int32_t index) const noexcept
Get the operation for an input tensor.
Definition: NvInfer.h:3651
void setOperation(int32_t index, MatrixOperation op) noexcept
Set the operation for an input tensor.
Definition: NvInfer.h:3639
A non-maximum suppression layer in a network definition.
Definition: NvInfer.h:6127
virtual ~INMSLayer() noexcept=default
void setTopKBoxLimit(int32_t limit) noexcept
Set the TopK box limit parameter for the layer.
Definition: NvInfer.h:6164
void setBoundingBoxFormat(BoundingBoxFormat fmt) noexcept
Set the bounding box format parameter for the layer.
Definition: NvInfer.h:6138
BoundingBoxFormat getBoundingBoxFormat() const noexcept
Get the bounding box format parameter for the layer.
Definition: NvInfer.h:6150
bool setIndicesType(DataType type) noexcept
Set the indices type for the layer.
Definition: NvInfer.h:6209
apiv::VNMSLayer * mImpl
Definition: NvInfer.h:6227
int32_t getTopKBoxLimit() const noexcept
Get the TopK box limit parameter for the layer.
Definition: NvInfer.h:6174
DataType getIndicesType() const noexcept
Return the NMS layer indices type.
Definition: NvInfer.h:6221
A network definition for input to the builder.
Definition: NvInfer.h:7054
IConcatenationLayer * addConcatenation(ITensor *const *inputs, int32_t nbInputs) noexcept
Add a concatenation layer to the network.
Definition: NvInfer.h:7282
IShuffleLayer * addShuffle(ITensor &input) noexcept
Add a shuffle layer to the network.
Definition: NvInfer.h:7345
INormalizationLayer * addNormalization(ITensor &input, ITensor &scale, ITensor &bias, uint32_t axesMask) noexcept
Add a normalization layer to the network.
Definition: NvInfer.h:8388
void setName(char const *name) noexcept
Sets the name of the network.
Definition: NvInfer.h:7773
ITopKLayer * addTopK(ITensor &input, TopKOperation op, int32_t k, uint32_t reduceAxes, DataType indicesType) noexcept
Add a TopK layer to the network.
Definition: NvInfer.h:7543
bool markDebug(ITensor &tensor) noexcept
Mark a tensor as a debug tensor.
Definition: NvInfer.h:7125
ILRNLayer * addLRN(ITensor &input, int64_t window, float alpha, float beta, float k) noexcept
Add a LRN layer to the network.
Definition: NvInfer.h:7226
ICumulativeLayer * addCumulative(ITensor &input, ITensor &axis, CumulativeOperation operation, bool exclusive, bool reverse) noexcept
Add a cumulative layer to the network.
Definition: NvInfer.h:8410
IAssertionLayer * addAssertion(ITensor &condition, char const *message) noexcept
Add an assertion layer to the network.
Definition: NvInfer.h:8075
TRT_DEPRECATED INonZeroLayer * addNonZero(ITensor &input) noexcept
Add a nonzero layer to the network.
Definition: NvInfer.h:7634
IConvolutionLayer * addConvolutionNd(ITensor &input, int64_t nbOutputMaps, Dims const &kernelSize, Weights kernelWeights, Weights biasWeights) noexcept
Add a multi-dimension convolution layer to the network.
Definition: NvInfer.h:7894
ICastLayer * addCast(ITensor &input, DataType toType) noexcept
Add a cast layer.
Definition: NvInfer.h:7703
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:7973
char const * getName() const noexcept
Returns the name associated with the network.
Definition: NvInfer.h:7787
IParametricReLULayer * addParametricReLU(ITensor &input, ITensor &slope) noexcept
Add a parametric ReLU layer to the network.
Definition: NvInfer.h:7872
ITensor * getOutput(int32_t index) const noexcept
Get the output tensor specified by the given index.
Definition: NvInfer.h:7446
ITensor * getInput(int32_t index) const noexcept
Get the input tensor specified by the given index.
Definition: NvInfer.h:7416
TRT_DEPRECATED ITopKLayer * addTopK(ITensor &input, TopKOperation op, int32_t k, uint32_t reduceAxes) noexcept
Add a TopK layer to the network.
Definition: NvInfer.h:7509
IDequantizeLayer * addDequantize(ITensor &input, ITensor &scale, DataType outputType) noexcept
Add a dequantization layer to the network.
Definition: NvInfer.h:8198
bool unmarkOutputForShapes(ITensor &tensor) noexcept
Undo markOutputForShapes.
Definition: NvInfer.h:7854
IFillLayer * addFill(Dims const &dimensions, FillOperation op, DataType outputType) noexcept
Add a fill layer to the network.
Definition: NvInfer.h:8101
ILoop * addLoop() noexcept
Add a loop to the network.
Definition: NvInfer.h:8004
IDynamicQuantizeLayer * addDynamicQuantize(ITensor &input, int32_t axis, int32_t blockSize, DataType outputType, DataType scaleType) noexcept
Add a dynamic quantization layer to the network.
Definition: NvInfer.h:8269
bool markUnfusedTensorsAsDebugTensors() noexcept
Mark unfused tensors as debug tensors.
Definition: NvInfer.h:7173
IActivationLayer * addActivation(ITensor &input, ActivationType type) noexcept
Add an activation layer to the network.
Definition: NvInfer.h:7207
ISliceLayer * addSlice(ITensor &input, Dims const &start, Dims const &size, Dims const &stride) noexcept
Add a slice layer to the network.
Definition: NvInfer.h:7749
virtual ~INetworkDefinition() noexcept=default
virtual IBuilder & getBuilder() const noexcept
Return the builder from which this INetworkDefinition was created.
Definition: NvInfer.h:8449
ILayer * getLayer(int32_t index) const noexcept
Get the layer specified by the given index.
Definition: NvInfer.h:7388
bool isDebugTensor(ITensor const &tensor) const noexcept
Check if a tensor is marked as debug tensor.
Definition: NvInfer.h:7151
bool getFlag(NetworkDefinitionCreationFlag networkDefinitionCreationFlag) const noexcept
Returns true if the network definition creation flag is set.
Definition: NvInfer.h:7825
IIfConditional * addIfConditional() noexcept
Add an if-then-else to the network.
Definition: NvInfer.h:8019
IErrorRecorder * getErrorRecorder() const noexcept
get the ErrorRecorder assigned to this interface.
Definition: NvInfer.h:8175
ISqueezeLayer * addSqueeze(ITensor &input, ITensor &axes) noexcept
Add a squeeze layer to the network.
Definition: NvInfer.h:8506
TRT_DEPRECATED INMSLayer * addNMS(ITensor &boxes, ITensor &scores, ITensor &maxOutputBoxesPerClass) noexcept
Add a non-maximum suppression layer to the network.
Definition: NvInfer.h:8325
IReverseSequenceLayer * addReverseSequence(ITensor &input, ITensor &sequenceLens) noexcept
Add a ReverseSequence layer to the network.
Definition: NvInfer.h:8362
int32_t getNbInputs() const noexcept
Get the number of inputs in the network.
Definition: NvInfer.h:7400
NetworkDefinitionCreationFlags getFlags() const noexcept
Get the network definition creation flags for this network definition object. Defaults to 0.
Definition: NvInfer.h:7813
IQuantizeLayer * addQuantize(ITensor &input, ITensor &scale, DataType outputType) noexcept
Add a quantization layer to the network.
Definition: NvInfer.h:8242
IReduceLayer * addReduce(ITensor &input, ReduceOperation operation, uint32_t reduceAxes, bool keepDimensions) noexcept
Add a reduce layer to the network.
Definition: NvInfer.h:7472
IUnaryLayer * addUnary(ITensor &input, UnaryOperation operation) noexcept
Add a unary layer to the network.
Definition: NvInfer.h:7331
IGridSampleLayer * addGridSample(ITensor &input, ITensor &grid) noexcept
Add a GridSample layer to the network.
Definition: NvInfer.h:8303
void removeTensor(ITensor &tensor) noexcept
remove a tensor from the network definition.
Definition: NvInfer.h:7718
bool areWeightsMarkedRefittable(char const *name) const noexcept
Whether the weight has been marked as refittable.
Definition: NvInfer.h:8487
ISelectLayer * addSelect(ITensor &condition, ITensor &thenInput, ITensor &elseInput) noexcept
Add a select layer to the network.
Definition: NvInfer.h:8058
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:8218
int32_t getNbLayers() const noexcept
Get the number of layers in the network.
Definition: NvInfer.h:7374
apiv::VNetworkDefinition * mImpl
Definition: NvInfer.h:8533
bool markOutputForShapes(ITensor &tensor) noexcept
Enable tensor's value to be computed by IExecutionContext::getShapeBinding.
Definition: NvInfer.h:7842
IOneHotLayer * addOneHot(ITensor &indices, ITensor &values, ITensor &depth, int32_t axis) noexcept
Add a OneHot layer to the network.
Definition: NvInfer.h:7362
IScaleLayer * addScale(ITensor &input, ScaleMode mode, Weights shift, Weights scale, Weights power) noexcept
Add a Scale layer to the network.
Definition: NvInfer.h:7252
void unmarkOutput(ITensor &tensor) noexcept
unmark a tensor as a network output.
Definition: NvInfer.h:7730
IIdentityLayer * addIdentity(ITensor &input) noexcept
Add an identity layer.
Definition: NvInfer.h:7688
IGatherLayer * addGatherV2(ITensor &data, ITensor &indices, GatherMode mode) noexcept
Add gather with specified mode, axis=0 and nbElementWiseDims=0.
Definition: NvInfer.h:7575
INonZeroLayer * addNonZero(ITensor &input, DataType indicesType) noexcept
Add a nonzero layer to the network.
Definition: NvInfer.h:7650
IElementWiseLayer * addElementWise(ITensor &input1, ITensor &input2, ElementWiseOperation op) noexcept
Add an elementwise layer to the network.
Definition: NvInfer.h:7309
IConstantLayer * addConstant(Dims const &dimensions, Weights weights) noexcept
Add a constant layer to the network.
Definition: NvInfer.h:7674
void setErrorRecorder(IErrorRecorder *recorder) noexcept
Set the ErrorRecorder for this interface.
Definition: NvInfer.h:8160
IPoolingLayer * addPoolingNd(ITensor &input, PoolingType type, Dims const &windowSize) noexcept
Add a multi-dimension pooling layer to the network.
Definition: NvInfer.h:7914
INMSLayer * addNMS(ITensor &boxes, ITensor &scores, ITensor &maxOutputBoxesPerClass, DataType indicesType) noexcept
Add a non-maximum suppression layer to the network.
Definition: NvInfer.h:8345
IRaggedSoftMaxLayer * addRaggedSoftMax(ITensor &input, ITensor &bounds) noexcept
Add a RaggedSoftMax layer to the network.
Definition: NvInfer.h:7594
IShapeLayer * addShape(ITensor &input) noexcept
Add a shape layer to the network.
Definition: NvInfer.h:7803
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:7559
IAttention * addAttention(ITensor &query, ITensor &key, ITensor &value, AttentionNormalizationOp normOp, bool causal) noexcept
Add an attention to the network.
Definition: NvInfer.h:8437
bool unmarkWeightsRefittable(char const *name) noexcept
Unmark weights as refittable when the builder flag kREFIT_INDIVIDUAL is set.
Definition: NvInfer.h:8474
bool markWeightsRefittable(char const *name) noexcept
Mark weights as refittable when the builder flag kREFIT_INDIVIDUAL is set.
Definition: NvInfer.h:8462
IDeconvolutionLayer * addDeconvolutionNd(ITensor &input, int64_t nbOutputMaps, Dims kernelSize, Weights kernelWeights, Weights biasWeights) noexcept
Add a multi-dimension deconvolution layer to the network.
Definition: NvInfer.h:7936
IResizeLayer * addResize(ITensor &input) noexcept
Add a resize layer to the network.
Definition: NvInfer.h:7990
IUnsqueezeLayer * addUnsqueeze(ITensor &input, ITensor &axes) noexcept
Add an unsqueeze layer to the network.
Definition: NvInfer.h:8527
IMatrixMultiplyLayer * addMatrixMultiply(ITensor &input0, MatrixOperation op0, ITensor &input1, MatrixOperation op1) noexcept
Add a MatrixMultiply layer to the network.
Definition: NvInfer.h:7615
ISoftMaxLayer * addSoftMax(ITensor &input) noexcept
Add a SoftMax layer to the network.
Definition: NvInfer.h:7265
bool unmarkDebug(ITensor &tensor) noexcept
Unmark a tensor as a debug tensor.
Definition: NvInfer.h:7141
IEinsumLayer * addEinsum(ITensor *const *inputs, int32_t nbInputs, char const *equation) noexcept
Add an Einsum layer to the network.
Definition: NvInfer.h:8285
void markOutput(ITensor &tensor) noexcept
Mark a tensor as a network output.
Definition: NvInfer.h:7107
IPaddingLayer * addPaddingNd(ITensor &input, Dims const &prePadding, Dims const &postPadding) noexcept
Add a padding layer to the network. Only 2D padding is currently supported.
Definition: NvInfer.h:8117
int32_t getNbOutputs() const noexcept
Get the number of outputs in the network.
Definition: NvInfer.h:7430
bool setWeightsName(Weights weights, char const *name) noexcept
Associate a name with all current uses of the given weights.
Definition: NvInfer.h:8141
bool unmarkUnfusedTensorsAsDebugTensors() noexcept
Undo the marking of unfused tensors as debug tensors.
Definition: NvInfer.h:7187
Forward declaration of IEngineInspector for use by other interfaces.
Definition: NvInferRuntime.h:51
Definition: NvInfer.h:3683
DataType getIndicesType() const noexcept
Return the NonZero layer indices type.
Definition: NvInfer.h:3707
bool setIndicesType(DataType type) noexcept
Set the indices type for the layer.
Definition: NvInfer.h:3695
virtual ~INonZeroLayer() noexcept=default
A normalization layer in a network definition.
Definition: NvInfer.h:6316
float getEpsilon() const noexcept
Get the epsilon value used for the normalization calculation.
Definition: NvInfer.h:6335
uint32_t getAxes() const noexcept
Get the axes value used for the normalization calculation.
Definition: NvInfer.h:6355
virtual ~INormalizationLayer() noexcept=default
void setEpsilon(float eps) noexcept
Set the epsilon value used for the normalization calculation.
Definition: NvInfer.h:6325
DataType getComputePrecision() const noexcept
Get the compute precision of this layer.
Definition: NvInfer.h:6422
apiv::VNormalizationLayer * mImpl
Definition: NvInfer.h:6428
int64_t getNbGroups() const noexcept
Get the number of groups used to split the channels for the normalization calculation.
Definition: NvInfer.h:6386
void setAxes(uint32_t axesMask) noexcept
Set the reduction axes for the normalization calculation.
Definition: NvInfer.h:6345
void setComputePrecision(DataType type) noexcept
Set the compute precision of this layer.
Definition: NvInfer.h:6412
void setNbGroups(int64_t nbGroups) noexcept
Set the number of groups used to split the channels in the normalization calculation.
Definition: NvInfer.h:6376
A OneHot layer in a network definition.
Definition: NvInfer.h:5938
virtual ~IOneHotLayer() noexcept=default
apiv::VOneHotLayer * mImpl
Definition: NvInfer.h:5959
void setAxis(int32_t axis) noexcept
Set the axis parameter.
Definition: NvInfer.h:5945
int32_t getAxis() const noexcept
Get the value of the axis parameter.
Definition: NvInfer.h:5953
Optimization profile for dynamic input dimensions and shape tensors.
Definition: NvInferRuntime.h:2672
Layer that represents a padding operation.
Definition: NvInfer.h:2880
Dims getPostPaddingNd() const noexcept
Get the padding that is applied at the end of the tensor.
Definition: NvInfer.h:2929
void setPrePaddingNd(Dims const &padding) noexcept
Set the padding that is applied at the start of the tensor.
Definition: NvInfer.h:2891
virtual ~IPaddingLayer() noexcept=default
void setPostPaddingNd(Dims const &padding) noexcept
Set the padding that is applied at the end of the tensor.
Definition: NvInfer.h:2917
Dims getPrePaddingNd() const noexcept
Get the padding that is applied at the start of the tensor.
Definition: NvInfer.h:2903
apiv::VPaddingLayer * mImpl
Definition: NvInfer.h:2935
Layer that represents a parametric ReLU operation.
Definition: NvInfer.h:3898
apiv::VParametricReLULayer * mImpl
Definition: NvInfer.h:3900
virtual ~IParametricReLULayer() noexcept=default
Single registration point for all plugins in an application. It is used to find plugin implementation...
Definition: NvInferRuntimeCommon.h:56
Plugin class for user-implemented layers.
Definition: NvInferRuntimePlugin.h:139
Layer type for pluginV2.
Definition: NvInfer.h:2617
virtual ~IPluginV2Layer() noexcept=default
apiv::VPluginV2Layer * mImpl
Definition: NvInfer.h:2630
IPluginV2 & getPlugin() noexcept
Get the plugin for the layer.
Definition: NvInfer.h:2624
Layer type for V3 plugins.
Definition: NvInfer.h:2644
virtual ~IPluginV3Layer() noexcept=default
IPluginV3 & getPlugin() noexcept
Get the plugin for the layer.
Definition: NvInfer.h:2651
apiv::VPluginV3Layer * mImpl
Definition: NvInfer.h:2657
A Pooling layer in a network definition.
Definition: NvInfer.h:1379
PoolingType getPoolingType() const noexcept
Get the type of activation to be performed.
Definition: NvInfer.h:1398
PaddingMode getPaddingMode() const noexcept
Get the padding mode.
Definition: NvInfer.h:1531
Dims getPostPadding() const noexcept
Get the padding.
Definition: NvInfer.h:1507
bool getAverageCountExcludesPadding() const noexcept
Get whether average pooling uses as a denominator the overlap area between the window and the unpadde...
Definition: NvInfer.h:1451
Dims getPrePadding() const noexcept
Get the pre-padding.
Definition: NvInfer.h:1479
void setPoolingType(PoolingType type) noexcept
Set the type of activation to be performed.
Definition: NvInfer.h:1388
void setWindowSizeNd(Dims const &windowSize) noexcept
Set the multi-dimension window size for pooling.
Definition: NvInfer.h:1544
void setPaddingMode(PaddingMode paddingMode) noexcept
Set the padding mode.
Definition: NvInfer.h:1520
Dims getWindowSizeNd() const noexcept
Get the multi-dimension window size for pooling.
Definition: NvInfer.h:1554
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:1440
void setPaddingNd(Dims const &padding) noexcept
Set the multi-dimension padding for pooling.
Definition: NvInfer.h:1598
float getBlendFactor() const noexcept
Get the blending factor for the max_average_blend mode: max_average_blendPool = (1-blendFactor)*maxPo...
Definition: NvInfer.h:1426
void setStrideNd(Dims const &stride) noexcept
Set the multi-dimension stride for pooling.
Definition: NvInfer.h:1569
Dims getStrideNd() const noexcept
Get the multi-dimension stride for pooling.
Definition: NvInfer.h:1579
virtual ~IPoolingLayer() noexcept=default
Dims getPaddingNd() const noexcept
Get the multi-dimension padding for pooling.
Definition: NvInfer.h:1610
void setPostPadding(Dims const &padding) noexcept
Set the multi-dimension post-padding for pooling.
Definition: NvInfer.h:1497
void setPrePadding(Dims const &padding) noexcept
Set the multi-dimension pre-padding for pooling.
Definition: NvInfer.h:1469
void setBlendFactor(float blendFactor) noexcept
Set the blending factor for the max_average_blend mode: max_average_blendPool = (1-blendFactor)*maxPo...
Definition: NvInfer.h:1413
A Quantize layer in a network definition.
Definition: NvInfer.h:5391
void setToType(DataType toType) noexcept
Set the Quantize layer output type.
Definition: NvInfer.h:5428
void setAxis(int32_t axis) noexcept
Set the quantization axis.
Definition: NvInfer.h:5412
int32_t getAxis() const noexcept
Get the quantization axis.
Definition: NvInfer.h:5401
virtual ~IQuantizeLayer() noexcept=default
DataType getToType() const noexcept
Return the Quantize layer output type.
Definition: NvInfer.h:5440
A RaggedSoftmax layer in a network definition.
Definition: NvInfer.h:3732
apiv::VRaggedSoftMaxLayer * mImpl
Definition: NvInfer.h:3734
virtual ~IRaggedSoftMaxLayer() noexcept=default
A recurrence layer in a network definition.
Definition: NvInfer.h:4604
virtual ~IRecurrenceLayer() noexcept=default
Layer that represents a reduction across a non-bool tensor.
Definition: NvInfer.h:2800
void setKeepDimensions(bool keepDimensions) noexcept
Set the boolean that specifies whether or not to keep the reduced dimensions for the layer.
Definition: NvInfer.h:2847
void setOperation(ReduceOperation op) noexcept
Set the reduce operation for the layer.
Definition: NvInfer.h:2807
ReduceOperation getOperation() const noexcept
Get the reduce operation for the layer.
Definition: NvInfer.h:2817
virtual ~IReduceLayer() noexcept=default
uint32_t getReduceAxes() const noexcept
Get the axes over which to reduce for the layer.
Definition: NvInfer.h:2837
void setReduceAxes(uint32_t reduceAxes) noexcept
Set the axes over which to reduce.
Definition: NvInfer.h:2827
apiv::VReduceLayer * mImpl
Definition: NvInfer.h:2863
bool getKeepDimensions() const noexcept
Get the boolean that specifies whether or not to keep the reduced dimensions for the layer.
Definition: NvInfer.h:2857
A resize layer in a network definition.
Definition: NvInfer.h:4087
void setSelectorForSinglePixel(ResizeSelector selector) noexcept
Set coordinate selector function when resized to single pixel.
Definition: NvInfer.h:4248
void setNearestRounding(ResizeRoundMode value) noexcept
Set rounding mode for nearest neighbor resize.
Definition: NvInfer.h:4272
virtual ~IResizeLayer() noexcept=default
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:4166
void setOutputDimensions(Dims const &dimensions) noexcept
Set the output dimensions.
Definition: NvInfer.h:4107
void setCubicCoeff(float A) noexcept
Set the coefficient 'A' used in cubic interpolation.
Definition: NvInfer.h:4304
void setScales(float const *scales, int32_t nbScales) noexcept
Set the resize scales.
Definition: NvInfer.h:4147
float getCubicCoeff() const noexcept
Get the coefficient 'A' used in cubic interpolation.
Definition: NvInfer.h:4314
ResizeSelector getSelectorForSinglePixel() const noexcept
Get the coordinate selector function when resized to single pixel.
Definition: NvInfer.h:4258
InterpolationMode getResizeMode() const noexcept
Get resize mode for an input tensor.
Definition: NvInfer.h:4188
void setCoordinateTransformation(ResizeCoordinateTransformation coordTransform) noexcept
Set coordinate transformation function.
Definition: NvInfer.h:4223
void setExcludeOutside(bool excludeFlag) noexcept
Set the state for excluding outside pixels.
Definition: NvInfer.h:4327
void setResizeMode(InterpolationMode interpolationMode) noexcept
Set resize mode for an input tensor.
Definition: NvInfer.h:4178
Dims getOutputDimensions() const noexcept
Get the output dimensions.
Definition: NvInfer.h:4117
ResizeRoundMode getNearestRounding() const noexcept
Get rounding mode for nearest neighbor resize.
Definition: NvInfer.h:4282
bool getExcludeOutside() const noexcept
Get the state for excluding outside pixels.
Definition: NvInfer.h:4337
ResizeCoordinateTransformation getCoordinateTransformation() const noexcept
Get coordinate transformation function.
Definition: NvInfer.h:4233
A ReverseSequence layer in a network definition.
Definition: NvInfer.h:6244
void setSequenceAxis(int32_t sequenceAxis) noexcept
Set the sequence axis. Default is 0.
Definition: NvInfer.h:6277
int32_t getBatchAxis() const noexcept
Return the batch axis. Return 1 if no batch axis was set.
Definition: NvInfer.h:6264
apiv::VReverseSequenceLayer * mImpl
Definition: NvInfer.h:6293
int32_t getSequenceAxis() const noexcept
Return the sequence axis. Return 0 if no sequence axis was set.
Definition: NvInfer.h:6287
void setBatchAxis(int32_t batchAxis) noexcept
Set the batch axis. Default is 1.
Definition: NvInfer.h:6254
virtual ~IReverseSequenceLayer() noexcept=default
A Scale layer in a network definition.
Definition: NvInfer.h:1776
Weights getScale() const noexcept
Get the scale value.
Definition: NvInfer.h:1833
Weights getPower() const noexcept
Get the power value.
Definition: NvInfer.h:1853
void setScale(Weights scale) noexcept
Set the scale value.
Definition: NvInfer.h:1823
void setPower(Weights power) noexcept
Set the power value.
Definition: NvInfer.h:1843
ScaleMode getMode() const noexcept
Get the scale mode.
Definition: NvInfer.h:1793
void setShift(Weights shift) noexcept
Set the shift value.
Definition: NvInfer.h:1803
void setChannelAxis(int32_t channelAxis) noexcept
Set the channel axis.
Definition: NvInfer.h:1889
Weights getShift() const noexcept
Get the shift value.
Definition: NvInfer.h:1813
virtual ~IScaleLayer() noexcept=default
void setMode(ScaleMode mode) noexcept
Set the scale mode.
Definition: NvInfer.h:1783
int32_t getChannelAxis() const noexcept
Get the channel axis.
Definition: NvInfer.h:1868
A scatter layer in a network definition. Supports several kinds of scattering.
Definition: NvInfer.h:5866
void setMode(ScatterMode mode) noexcept
Set the scatter mode.
Definition: NvInfer.h:5873
apiv::VScatterLayer * mImpl
Definition: NvInfer.h:5907
void setAxis(int32_t axis) noexcept
Set the axis used by ScatterMode::kELEMENTS.
Definition: NvInfer.h:5893
int32_t getAxis() const noexcept
Get the axis.
Definition: NvInfer.h:5901
ScatterMode getMode() const noexcept
Get the scatter mode.
Definition: NvInfer.h:5883
virtual ~IScatterLayer() noexcept=default
Select elements from two data tensors based on a condition tensor.
Definition: NvInfer.h:4912
virtual ~ISelectLayer() noexcept=default
Layer type for getting shape of a tensor.
Definition: NvInfer.h:3405
virtual ~IShapeLayer() noexcept=default
apiv::VShapeLayer * mImpl
Definition: NvInfer.h:3407
Layer type for shuffling data.
Definition: NvInfer.h:2968
apiv::VShuffleLayer * mImpl
Definition: NvInfer.h:3126
void setFirstTranspose(Permutation permutation) noexcept
Set the permutation applied by the first transpose operation.
Definition: NvInfer.h:2979
void setSecondTranspose(Permutation permutation) noexcept
Set the permutation applied by the second transpose operation.
Definition: NvInfer.h:3079
Dims getReshapeDimensions() const noexcept
Get the reshaped dimensions.
Definition: NvInfer.h:3032
void setReshapeDimensions(Dims const &dimensions) noexcept
Set the reshaped dimensions.
Definition: NvInfer.h:3019
Permutation getFirstTranspose() const noexcept
Get the permutation applied by the first transpose operation.
Definition: NvInfer.h:2991
virtual ~IShuffleLayer() noexcept=default
Permutation getSecondTranspose() const noexcept
Get the permutation applied by the second transpose operation.
Definition: NvInfer.h:3091
bool getZeroIsPlaceholder() const noexcept
Get meaning of 0 in reshape dimensions.
Definition: NvInfer.h:3120
void setZeroIsPlaceholder(bool zeroIsPlaceholder) noexcept
Set meaning of 0 in reshape dimensions.
Definition: NvInfer.h:3107
Slices an input tensor into an output tensor based on the offset and strides.
Definition: NvInfer.h:3220
void setStride(Dims const &stride) noexcept
Set the stride for computing the output slice data.
Definition: NvInfer.h:3289
apiv::VSliceLayer * mImpl
Definition: NvInfer.h:3388
virtual ~ISliceLayer() noexcept=default
void setSize(Dims const &size) noexcept
Set the dimensions of the output slice.
Definition: NvInfer.h:3260
void setAxes(Dims const &axes) noexcept
Set the axes for this ISliceLayer.
Definition: NvInfer.h:3367
void setStart(Dims const &start) noexcept
Set the start offset that the slice layer uses to create the output slice.
Definition: NvInfer.h:3231
Dims getStart() const noexcept
Get the start offset for the slice layer.
Definition: NvInfer.h:3246
void setMode(SampleMode mode) noexcept
Set the slice mode.
Definition: NvInfer.h:3314
Dims getSize() const noexcept
Get dimensions of the output slice.
Definition: NvInfer.h:3275
SampleMode getMode() const noexcept
Get the slice mode.
Definition: NvInfer.h:3324
Dims getStride() const noexcept
Get the stride for the output slice.
Definition: NvInfer.h:3304
Dims getAxes() const noexcept
Get the axes for this ISliceLayer.
Definition: NvInfer.h:3382
A Softmax layer in a network definition.
Definition: NvInfer.h:1920
void setAxes(uint32_t axes) noexcept
Set the axis along which softmax is computed. Currently, only one axis can be set.
Definition: NvInfer.h:1942
uint32_t getAxes() const noexcept
Get the axis along which softmax occurs.
Definition: NvInfer.h:1952
virtual ~ISoftMaxLayer() noexcept=default
Layer that represents a squeeze operation, removing unit dimensions of the input tensor on a set of a...
Definition: NvInfer.h:6442
virtual ~ISqueezeLayer() noexcept=default
apiv::VSqueezeLayer * mImpl
Definition: NvInfer.h:6459
A tensor in a network definition.
Definition: NvInfer.h:185
void setAllowedFormats(TensorFormats formats) noexcept
Set allowed formats for an input or output tensor. By default all formats are allowed....
Definition: NvInfer.h:336
void setDimensions(Dims const &dimensions) noexcept
Set the dimensions of a tensor.
Definition: NvInfer.h:233
void setName(char const *name) noexcept
Set the tensor name.
Definition: NvInfer.h:202
bool isExecutionTensor() const noexcept
Whether the tensor is an execution tensor.
Definition: NvInfer.h:401
char const * getName() const noexcept
Get the tensor name.
Definition: NvInfer.h:214
bool isShapeTensor() const noexcept
Whether the tensor is a shape tensor.
Definition: NvInfer.h:380
bool isNetworkInput() const noexcept
Whether the tensor is a network input.
Definition: NvInfer.h:306
TRT_DEPRECATED void setType(DataType type) noexcept
Set the data type of a tensor.
Definition: NvInfer.h:283
bool isNetworkOutput() const noexcept
Whether the tensor is a network output.
Definition: NvInfer.h:314
DataType getType() const noexcept
Get the data type of a tensor.
Definition: NvInfer.h:298
apiv::VTensor * mImpl
Definition: NvInfer.h:448
virtual ~ITensor() noexcept=default
void setDimensionName(int32_t index, char const *name) noexcept
Name a dimension of an input tensor.
Definition: NvInfer.h:427
char const * getDimensionName(int32_t index) const noexcept
Get the name of an input dimension.
Definition: NvInfer.h:442
Dims getDimensions() const noexcept
Get the dimensions of a tensor.
Definition: NvInfer.h:247
TensorFormats getAllowedFormats() const noexcept
Get a bitmask of TensorFormat values that the tensor supports. For a shape tensor,...
Definition: NvInfer.h:349
Class to handle tactic timing info collected from builder.
Definition: NvInfer.h:8855
int64_t queryKeys(TimingCacheKey *keyBuffer, int64_t capacity) const noexcept
Query cache keys from Timing Cache.
Definition: NvInfer.h:8921
bool combine(ITimingCache const &inputCache, bool ignoreMismatch) noexcept
Combine input timing cache into local instance.
Definition: NvInfer.h:8892
TimingCacheValue query(TimingCacheKey const &key) const noexcept
Query value in a cache entry.
Definition: NvInfer.h:8938
virtual ~ITimingCache() noexcept=default
bool update(TimingCacheKey const &key, TimingCacheValue const &value) noexcept
Update values in a cache entry.
Definition: NvInfer.h:8960
apiv::VTimingCache * mImpl
Definition: NvInfer.h:8966
bool reset() noexcept
Empty the timing cache.
Definition: NvInfer.h:8902
Layer that represents a TopK reduction.
Definition: NvInfer.h:3445
void setK(int32_t k) noexcept
Set the static k value for the layer.
Definition: NvInfer.h:3476
void setReduceAxes(uint32_t reduceAxes) noexcept
Set which axes to reduce for the layer.
Definition: NvInfer.h:3500
TopKOperation getOperation() const noexcept
Get the operation for the layer.
Definition: NvInfer.h:3462
apiv::VTopKLayer * mImpl
Definition: NvInfer.h:3559
void setOperation(TopKOperation op) noexcept
Set the operation for the layer.
Definition: NvInfer.h:3452
bool setIndicesType(DataType type) noexcept
Set the indices type for the layer.
Definition: NvInfer.h:3541
int32_t getK() const noexcept
Get the k value for the layer.
Definition: NvInfer.h:3490
uint32_t getReduceAxes() const noexcept
Get the axes to reduce for the layer.
Definition: NvInfer.h:3510
virtual ~ITopKLayer() noexcept=default
DataType getIndicesType() const noexcept
Return the TopK layer indices type.
Definition: NvInfer.h:3553
A layer that represents a trip-count limiter.
Definition: NvInfer.h:4725
TripLimit getTripLimit() const noexcept
Get a trip limiter type.
Definition: NvInfer.h:4730
virtual ~ITripLimitLayer() noexcept=default
Layer that represents an unary operation.
Definition: NvInfer.h:2725
void setOperation(UnaryOperation op) noexcept
Set the unary operation for the layer.
Definition: NvInfer.h:2734
apiv::VUnaryLayer * mImpl
Definition: NvInfer.h:2750
UnaryOperation getOperation() const noexcept
Get the unary operation for the layer.
Definition: NvInfer.h:2744
virtual ~IUnaryLayer() noexcept=default
Layer that represents an unsqueeze operation, which reshapes the input tensor by inserting unit-lengt...
Definition: NvInfer.h:6471
virtual ~IUnsqueezeLayer() noexcept=default
apiv::VUnsqueezeLayer * mImpl
Definition: NvInfer.h:6488
An Interface class for version control.
Definition: NvInferRuntimeBase.h:278
Version information associated with a TRT interface.
Definition: NvInferRuntimeBase.h:243
An array of weights used as a layer parameter.
Definition: NvInferRuntime.h:124
Definition: NvInferRuntimeBase.h:415
Definition: NvInferRuntime.h:1656
Definition: NvInferPluginBase.h:206
Definition: NvInfer.h:9204
virtual bool stepComplete(char const *phaseName, int32_t step) noexcept=0
Signal that a step of an optimizer phase has finished.
virtual ~IProgressMonitor() noexcept=default
IProgressMonitor()=default
virtual void phaseFinish(char const *phaseName) noexcept=0
Signal that a phase of the optimizer has finished.
virtual void phaseStart(char const *phaseName, char const *parentPhase, int32_t nbSteps) noexcept=0
Signal that a phase of the optimizer has started.
Definition: NvInferRuntime.h:666
IBuilder * createInferBuilder(ILogger &logger) noexcept
Create an instance of an IBuilder class.
Definition: NvInfer.h:10468
The TensorRT-RTX API version 1 namespace.
uint32_t TacticSources
Represents a collection of one or more TacticSource values combine using bitwise-OR operations.
Definition: NvInferRuntime.h:2958
ResizeSelector
The coordinate selector when resize to single pixel output.
Definition: NvInfer.h:3992
@ kFORMULA
Use formula to map the original index.
@ kUPPER
Select the upper left pixel.
EngineCapability
List of supported engine capability flows.
Definition: NvInferRuntime.h:76
MemoryPoolType
The type for memory pools used by TensorRT.
Definition: NvInfer.h:8977
ScaleMode
Controls how shift, scale and power are applied in a Scale layer.
Definition: NvInfer.h:1733
@ kUNIFORM
Identical coefficients across all elements of the tensor.
@ kCHANNEL
Per-channel coefficients.
RuntimePlatform
Describes the intended runtime platform (operating system and CPU architecture) for the execution of ...
Definition: NvInfer.h:8554
HardwareCompatibilityLevel
Describes requirements of compatibility with GPU architectures other than that of the GPU on which th...
Definition: NvInfer.h:9096
@ kSAME_COMPUTE_CAPABILITY
CumulativeOperation
Enumerates the cumulative operations that may be performed by a Cumulative layer.
Definition: NvInfer.h:6504
BoundingBoxFormat
Representation of bounding box data used for the Boxes input tensor in INMSLayer.
Definition: NvInfer.h:6058
@ kCENTER_SIZES
(x_center, y_center, width, height) where (x_center, y_center) is the center point of the box
@ kCORNER_PAIRS
(x1, y1, x2, y2) where (x1, y1) and (x2, y2) are any pair of diagonal corners
constexpr int32_t EnumMax< BuilderFlag >() noexcept
Definition: NvInfer.h:8789
constexpr int32_t EnumMax< LayerType >() noexcept
Definition: NvInfer.h:119
ComputeCapability
Describes compute capability that an engine will be built for.
Definition: NvInfer.h:9145
@ kSM120
Target NVIDIA Blackwell GPU architecture (SM 12.0).
@ kSM75
Target NVIDIA Turing GPU architecture (SM 7.5).
@ kSM80
Target NVIDIA Ampere GPU architecture (SM 8.0).
@ kCURRENT
Use the compute capability of the current GPU in the environment.
@ kSM89
Target NVIDIA Ada Lovelace GPU architecture (SM 8.9).
@ kSM86
Target NVIDIA Ampere GPU architecture (SM 8.6).
UnaryOperation
Enumerates the unary operations that may be performed by a Unary layer.
Definition: NvInfer.h:2678
@ kISINF
Return true if input value equals +/- infinity for floating-point data type.
@ kCOSH
Hyperbolic cosine.
@ kACOSH
Inverse hyperbolic cosine.
@ kERF
Gauss error function.
@ kISNAN
Return true if input value is a NaN for floating-point data type.
@ kROUND
Round to nearest even for floating-point data type.
@ kATANH
Inverse hyperbolic tangent.
@ kASINH
Inverse hyperbolic sine.
@ kSIGN
Sign, If input > 0, output 1; if input < 0, output -1; if input == 0, output 0.
constexpr int32_t EnumMax< ReduceOperation >() noexcept
Definition: NvInfer.h:2787
constexpr int32_t EnumMax< TripLimit >() noexcept
Definition: NvInfer.h:4393
ActivationType
Enumerates the types of activation to perform in an activation layer.
Definition: NvInfer.h:138
@ kSELU
Selu activation: x>0 ? beta * x : beta * (alpha*exp(x) - alpha)
@ kSCALED_TANH
Scaled tanh activation: alpha*tanh(beta*x)
@ kRELU
Rectified linear activation.
@ kELU
Elu activation: x>=0 ? x : alpha * (exp(x) - 1).
@ kLEAKY_RELU
LeakyRelu activation: x>=0 ? x : alpha * x.
@ kSOFTSIGN
Softsign activation: x / (1+|x|)
@ kHARD_SIGMOID
Hard sigmoid activation: max(0, min(1, alpha*x+beta))
@ kTHRESHOLDED_RELU
Thresholded ReLU activation: x>alpha ? x : 0.
@ kSIGMOID
Sigmoid activation.
@ kCLIP
Clip activation: max(alpha, min(beta, x))
@ kGELU_TANH
GELU tanh activation: 0.5 * x * (1 + tanh(sqrt(2/pi) * (0.044715F * pow(x, 3) + x)))
@ kGELU_ERF
GELU erf activation: 0.5 * x * (1 + erf(sqrt(0.5) * x))
@ kSOFTPLUS
Parametric softplus activation: alpha*log(exp(beta*x)+1)
FillOperation
Enumerates the tensor fill operations that may performed by a fill layer.
Definition: NvInfer.h:4973
@ kRANDOM_UNIFORM
Randomly draw values from a uniform distribution.
@ kRANDOM_NORMAL
Randomly draw values from a normal distribution.
ResizeRoundMode
The rounding mode for nearest neighbor resize.
Definition: NvInfer.h:4022
@ kHALF_DOWN
Round half down.
PaddingMode
Enumerates the modes of padding to perform in convolution, deconvolution and pooling layer,...
Definition: NvInfer.h:911
@ kSAME_LOWER
Use SAME padding, with prePadding >= postPadding.
@ kEXPLICIT_ROUND_DOWN
Use explicit padding, rounding output size down.
@ kEXPLICIT_ROUND_UP
Use explicit padding, rounding output size up.
@ kSAME_UPPER
Use SAME padding, with prePadding <= postPadding.
TripLimit
Enum that describes kinds of trip limits.
Definition: NvInfer.h:4381
@ kWHILE
Tensor is a scalar of type kBOOL. Loop terminates when value is false.
@ kCOUNT
Tensor is a scalar of type kINT32 or kINT64 that contains the trip count.
uint32_t NetworkDefinitionCreationFlags
Represents one or more NetworkDefinitionCreationFlag flags using binary OR operations....
Definition: NvInfer.h:10148
PreviewFeature
Define preview features.
Definition: NvInfer.h:9052
@ kRUNTIME_ACTIVATION_RESIZE_10_10
@ kALIASED_PLUGIN_IO_10_03
TilingOptimizationLevel
Define the optimization levels for Tiling.
Definition: NvInfer.h:9171
@ kFAST
Use a fast algorithm and heuristic based strategy. Slightly increases engine build time.
@ kFULL
Increase search space even wider. Significantly increases engine build time.
constexpr int32_t EnumMax< GatherMode >() noexcept
Definition: NvInfer.h:2437
DataType
The type of weights and tensors. The datatypes other than kBOOL, kINT32, and kINT64 are "activation d...
Definition: NvInferRuntimeBase.h:145
uint32_t BuilderFlags
Represents one or more BuilderFlag values using binary OR operations, e.g., 1U << BuilderFlag::kFP16 ...
Definition: NvInfer.h:8586
DeviceType
The device that this layer/network will execute on.
Definition: NvInferRuntime.h:1350
constexpr int32_t EnumMax< ScaleMode >() noexcept
Definition: NvInfer.h:1745
LayerType
The type values of layer classes.
Definition: NvInfer.h:57
@ kGRID_SAMPLE
Grid sample layer.
@ kRAGGED_SOFTMAX
Ragged softmax layer.
@ kDECONVOLUTION
Deconvolution layer.
@ kASSERTION
Assertion layer.
@ kATTENTION_INPUT
Attention Input.
@ kMATRIX_MULTIPLY
Matrix multiply layer.
@ kCONDITION
Condition layer.
@ kCUMULATIVE
Cumulative layer.
@ kCONDITIONAL_INPUT
Conditional Input layer.
@ kIDENTITY
Identity layer.
@ kNORMALIZATION
Normalization layer.
@ kQUANTIZE
Quantize layer.
@ kCONVOLUTION
Convolution layer.
@ kPARAMETRIC_RELU
Parametric ReLU layer.
@ kATTENTION_OUTPUT
Attention Output.
@ kUNSQUEEZE
Unsqueeze Layer.
@ kCONCATENATION
Concatenation layer.
@ kREVERSE_SEQUENCE
Reverse sequence layer.
@ kRECURRENCE
Loop Recurrence layer.
@ kDEQUANTIZE
Dequantize layer.
@ kPLUGIN_V3
PluginV3 layer.
@ kITERATOR
Loop Iterator layer.
@ kTRIP_LIMIT
Loop Trip limit layer.
@ kDYNAMIC_QUANTIZE
Dynamic Quantize layer.
@ kUNARY
UnaryOp operation Layer.
@ kACTIVATION
Activation layer.
@ kELEMENTWISE
Elementwise layer.
@ kPLUGIN_V2
PluginV2 layer.
@ kLOOP_OUTPUT
Loop output layer.
@ kCONDITIONAL_OUTPUT
Conditional Output layer.
@ kCONSTANT
Constant layer.
@ kNON_ZERO
NonZero layer.
SampleMode
Controls how ISliceLayer and IGridSample handle out-of-bounds coordinates.
Definition: NvInfer.h:3136
@ kCLAMP
Out of bounds indices are clamped to bounds.
@ kSTRICT_BOUNDS
Fail with error when the coordinates are out of bounds.
@ kWRAP
Coordinates wrap around periodically.
GatherMode
Control form of IGatherLayer.
Definition: NvInfer.h:2425
@ kDEFAULT
Similar to ONNX Gather.
@ kELEMENT
Similar to ONNX GatherElements.
@ kND
Similar to ONNX GatherND.
uint32_t TensorFormats
It is capable of representing one or more TensorFormat by binary OR operations, e....
Definition: NvInfer.h:130
ProfilingVerbosity
List of verbosity levels of layer information exposed in NVTX annotations and in IEngineInspector.
Definition: NvInferRuntime.h:2970
NetworkDefinitionCreationFlag
List of immutable network properties expressed at network creation time. NetworkDefinitionCreationFla...
Definition: NvInfer.h:10159
ElementWiseOperation
Enumerates the binary operations that may be performed by an ElementWise layer.
Definition: NvInfer.h:2335
@ kSUB
Subtract the second element from the first.
@ kSUM
Sum of the two elements.
@ kPROD
Product of the two elements.
@ kFLOOR_DIV
Floor division of the first element by the second.
@ kEQUAL
Check if two elements are equal.
@ kAND
Logical AND of two elements.
@ kOR
Logical OR of two elements.
@ kMIN
Minimum of the two elements.
@ kPOW
The first element to the power of the second element.
@ kLESS
Check if element in first tensor is less than corresponding element in second tensor.
@ kGREATER
Check if element in first tensor is greater than corresponding element in second tensor.
@ kXOR
Logical XOR of two elements.
@ kDIV
Divide the first element by the second.
constexpr int32_t EnumMax< SampleMode >() noexcept
Definition: NvInfer.h:3152
InterpolationMode
Enumerates various modes of interpolation.
Definition: NvInfer.h:3910
@ kNEAREST
ND (0 < N <= 8) nearest neighbor resizing.
@ kCUBIC
Supports bicubic (2D) interpolation.
@ kLINEAR
Supports linear (1D), bilinear (2D), and trilinear (3D) interpolation.
BuilderFlag
List of valid modes that the builder can enable when creating an engine from a network definition.
Definition: NvInfer.h:8596
@ kWEIGHT_STREAMING
Enable weight streaming for the current engine.
@ kDEBUG
Enable debugging of layers via synchronizing after every layer.
@ kGPU_FALLBACK
Enable layers marked to execute on GPU if layer cannot execute on DLA.
@ kSPARSE_WEIGHTS
Allow the builder to examine weights and use optimized functions when weights have suitable sparsity.
@ kERROR_ON_TIMING_CACHE_MISS
@ kEDITABLE_TIMING_CACHE
Enable editable timing cache.
@ kPREFER_PRECISION_CONSTRAINTS
@ kDISTRIBUTIVE_INDEPENDENCE
@ kSTRIP_PLAN
Strip the refittable weights from the engine plan file.
@ kMONITOR_MEMORY
Enable memory monitor during build time.
@ kDISABLE_TIMING_CACHE
Disable reuse of timing information across identical layers.
@ kDISABLE_COMPILATION_CACHE
@ kREFIT
Enable building a refittable engine.
@ kOBEY_PRECISION_CONSTRAINTS
@ kREJECT_EMPTY_ALGORITHMS
constexpr int32_t EnumMax< TopKOperation >() noexcept
Definition: NvInfer.h:3428
TENSORRTAPI nvinfer1::IPluginRegistry * getBuilderPluginRegistry(nvinfer1::EngineCapability capability) noexcept
Return the plugin registry for building a Standard engine, or nullptr if no registry exists.
constexpr int32_t EnumMax< MemoryPoolType >() noexcept
Definition: NvInfer.h:9038
TopKOperation
Enumerates the operations that may be performed by a TopK layer.
Definition: NvInfer.h:3417
ReduceOperation
Enumerates the reduce operations that may be performed by a Reduce layer.
Definition: NvInfer.h:2773
constexpr int32_t EnumMax< LoopOutput >() noexcept
Definition: NvInfer.h:4370
constexpr int32_t EnumMax< NetworkDefinitionCreationFlag >() noexcept
Definition: NvInfer.h:10178
ScatterMode
Control form of IScatterLayer.
Definition: NvInfer.h:5792
MatrixOperation
Enumerates the operations that may be performed on a tensor by IMatrixMultiplyLayer before multiplica...
Definition: NvInfer.h:3570
@ kTRANSPOSE
Like kNONE, but transpose the matrix dimensions.
ResizeCoordinateTransformation
The resize coordinate transformation function.
Definition: NvInfer.h:3938
constexpr int32_t EnumMax< UnaryOperation >() noexcept
Definition: NvInfer.h:2712
LoopOutput
Enum that describes kinds of loop outputs.
Definition: NvInfer.h:4353
@ kLAST_VALUE
Output value is value of tensor for last iteration.
@ kCONCATENATE
Output value is concatenation of values of tensor for each iteration, in forward order.
@ kREVERSE
Output value is concatenation of values of tensor for each iteration, in reverse order.
constexpr int32_t EnumMax< BoundingBoxFormat >() noexcept
Definition: NvInfer.h:6071
constexpr int32_t EnumMax< MatrixOperation >() noexcept
Definition: NvInfer.h:3598
PoolingType
The type of pooling to perform in a pooling layer.
Definition: NvInfer.h:1347
@ kAVERAGE
Average over elements. If the tensor is padded, the count includes the padding.
@ kMAX
Maximum over elements.
@ kMAX_AVERAGE_BLEND
Blending between max and average pooling: (1-blendFactor)*maxPool + blendFactor*avgPool.
v_1_0::IProgressMonitor IProgressMonitor
Definition: NvInfer.h:9287
constexpr int32_t EnumMax< FillOperation >() noexcept
Definition: NvInfer.h:5004
AttentionNormalizationOp
Enumerates the operations that may be performed by the normalization in the attention subgraph.
Definition: NvInfer.h:6639
constexpr int32_t EnumMax< ScatterMode >() noexcept
Definition: NvInfer.h:5803
Represents a permutation of dimensions.
Definition: NvInfer.h:2945
Declaration of EnumMaxImpl struct to store maximum number of elements in an enumeration type.
Definition: NvInferRuntimeBase.h:128
The key to retrieve timing cache entries.
Definition: NvInfer.h:8815
Definition: NvInfer.h:8829
uint64_t tacticHash
Hash of the selected tactic.
Definition: NvInfer.h:8831
float timingMSec
Timing of this tactic in milliseconds. Negative numbers and NaN are invalid values.
Definition: NvInfer.h:8833