mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
bca9a33b02
* Update TensorRT-LLM --------- Co-authored-by: Timur Abishev <abishev.timur@gmail.com> Co-authored-by: MahmoudAshraf97 <hassouna97.ma@gmail.com> Co-authored-by: Saeyoon Oh <saeyoon.oh@furiosa.ai> Co-authored-by: hattizai <hattizai@gmail.com>
322 lines
14 KiB
C++
322 lines
14 KiB
C++
/*
|
|
* Copyright (c) 2019-2024, NVIDIA CORPORATION. All rights reserved.
|
|
* Copyright (c) 2021, NAVER Corp. Authored by CLOVA.
|
|
*
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*/
|
|
|
|
#include "penaltyLayer.h"
|
|
#include "tensorrt_llm/common/cudaUtils.h"
|
|
#include "tensorrt_llm/kernels/penaltyKernels.h"
|
|
#include "tensorrt_llm/layers/defaultDecodingParams.h"
|
|
#include "tensorrt_llm/layers/layerUtils.h"
|
|
#include "tensorrt_llm/runtime/bufferManager.h"
|
|
|
|
#include <algorithm>
|
|
|
|
using namespace tensorrt_llm::common;
|
|
using namespace tensorrt_llm::kernels;
|
|
using namespace tensorrt_llm::runtime;
|
|
|
|
namespace tensorrt_llm::layers
|
|
{
|
|
|
|
template <typename T>
|
|
PenaltyLayer<T>::PenaltyLayer(executor::DecodingMode const& mode, DecoderDomain const& decoderDomain,
|
|
std::shared_ptr<BufferManager> bufferManager)
|
|
: BaseLayer(decoderDomain, bufferManager)
|
|
, mDecodingMode(mode)
|
|
{
|
|
TLLM_LOG_TRACE("%s start", __PRETTY_FUNCTION__);
|
|
|
|
initialize();
|
|
|
|
TLLM_LOG_TRACE("%s stop", __PRETTY_FUNCTION__);
|
|
}
|
|
|
|
template <typename T>
|
|
size_t PenaltyLayer<T>::getWorkspaceSize() const noexcept
|
|
{
|
|
return mPenaltyWorkspaceDevice->getSizeInBytes();
|
|
}
|
|
|
|
template <typename T>
|
|
void PenaltyLayer<T>::initialize()
|
|
{
|
|
TLLM_LOG_TRACE("%s start", __PRETTY_FUNCTION__);
|
|
|
|
allocateBuffer();
|
|
|
|
mCyclicStep = 0;
|
|
mRuntimeMaxSeqLen = 0;
|
|
mConfiguredBeamWidth = -1;
|
|
|
|
if (!mDecodingMode.isAuto())
|
|
{
|
|
mConfiguredBeamWidth = mDecoderDomain.getBeamWidth();
|
|
|
|
allocateWorkspace();
|
|
}
|
|
|
|
TLLM_LOG_TRACE("%s stop", __PRETTY_FUNCTION__);
|
|
}
|
|
|
|
template <typename T>
|
|
void PenaltyLayer<T>::allocateWorkspace()
|
|
{
|
|
TLLM_LOG_TRACE("%s start", __PRETTY_FUNCTION__);
|
|
|
|
if (mDecodingMode.isUseOccurrencePenalty())
|
|
{
|
|
|
|
auto const workspaceSize = mDecoderDomain.getBatchSize() * mDecoderDomain.getMaxDecodingTokens()
|
|
* mConfiguredBeamWidth * mDecoderDomain.getVocabSize();
|
|
mPenaltyWorkspaceDevice = mBufferManager->gpu(workspaceSize, nvinfer1::DataType::kINT32);
|
|
|
|
if (mDecodingMode.isBeamSearch())
|
|
{
|
|
mPenaltyWorkspacePrevDevice = mBufferManager->gpu(workspaceSize, nvinfer1::DataType::kINT32);
|
|
}
|
|
}
|
|
|
|
TLLM_LOG_TRACE("%s stop", __PRETTY_FUNCTION__);
|
|
}
|
|
|
|
template <typename T>
|
|
void PenaltyLayer<T>::allocateBuffer()
|
|
{
|
|
TLLM_LOG_TRACE("%s start", __PRETTY_FUNCTION__);
|
|
|
|
mLogitsPtrsHost = mBufferManager->pinnedPool(ITensor::makeShape({}), TRTDataType<T*>::value);
|
|
auto const batchSizeShape = ITensor::makeShape({mDecoderDomain.getBatchSize()});
|
|
mTemperature = mBufferManager->pinnedPool(batchSizeShape, TRTDataType<float>::value);
|
|
mRepetitionPenalty = mBufferManager->pinnedPool(batchSizeShape, TRTDataType<float>::value);
|
|
mPresencePenalty = mBufferManager->pinnedPool(batchSizeShape, TRTDataType<float>::value);
|
|
mFrequencyPenalty = mBufferManager->pinnedPool(batchSizeShape, TRTDataType<float>::value);
|
|
mMinLength = mBufferManager->pinnedPool(batchSizeShape, TRTDataType<SizeType32>::value);
|
|
|
|
if (mDecodingMode.isUseTemperature())
|
|
{
|
|
mTemperatureDevice = mBufferManager->gpu(batchSizeShape, nvinfer1::DataType::kFLOAT);
|
|
}
|
|
if (mDecodingMode.isUseRepetitionPenalty())
|
|
{
|
|
mRepetitionPenaltyDevice = mBufferManager->gpu(batchSizeShape, nvinfer1::DataType::kFLOAT);
|
|
}
|
|
if (mDecodingMode.isUsePresencePenalty())
|
|
{
|
|
mPresencePenaltyDevice = mBufferManager->gpu(batchSizeShape, nvinfer1::DataType::kFLOAT);
|
|
}
|
|
if (mDecodingMode.isUseFrequencyPenalty())
|
|
{
|
|
mFrequencyPenaltyDevice = mBufferManager->gpu(batchSizeShape, nvinfer1::DataType::kFLOAT);
|
|
}
|
|
if (mDecodingMode.isUseMinLength())
|
|
{
|
|
mMinLengthDevice = mBufferManager->gpu(batchSizeShape, nvinfer1::DataType::kINT32);
|
|
}
|
|
|
|
auto const runtimeLogitsDeviceSize = mDecoderDomain.getBatchSize() * mDecoderDomain.getMaxDecodingTokens()
|
|
* mDecoderDomain.getBeamWidth() * mDecoderDomain.getVocabSizePadded();
|
|
|
|
mRuntimeLogitsDevice = mBufferManager->gpu(ITensor::makeShape({runtimeLogitsDeviceSize}), TRTDataType<T>::value);
|
|
|
|
TLLM_LOG_TRACE("%s stop", __PRETTY_FUNCTION__);
|
|
}
|
|
|
|
template <typename T>
|
|
void PenaltyLayer<T>::setup(SizeType32 batchSize, SizeType32 beamWidth, BufferConstPtr batchSlots,
|
|
std::shared_ptr<BaseSetupParams> const& baseSetupParams)
|
|
{
|
|
TLLM_LOG_TRACE("%s start", __PRETTY_FUNCTION__);
|
|
|
|
auto setupParams = std::dynamic_pointer_cast<DynamicDecodeSetupParams>(baseSetupParams);
|
|
|
|
if (mConfiguredBeamWidth == -1)
|
|
{
|
|
// This code is left only for Python runtime
|
|
// In C++ runtime given maxBeamWidth should always be equal to the runtime beamWidth
|
|
TLLM_CHECK(mDecodingMode.isAuto());
|
|
mConfiguredBeamWidth = beamWidth;
|
|
mDecodingMode
|
|
= mConfiguredBeamWidth == 1 ? executor::DecodingMode::TopKTopP() : executor::DecodingMode::BeamSearch();
|
|
allocateWorkspace();
|
|
}
|
|
|
|
auto defaultBatchSlots = batchSlots ? batchSlots : getDefaultBatchSlots(batchSize, *mBufferManager);
|
|
|
|
// Setup penalties.
|
|
FillBuffers const fillBuffers{batchSize, mDecoderDomain.getBatchSize(), mBufferManager};
|
|
|
|
auto const& penaltyParams = setupParams->penaltyParams;
|
|
TLLM_CHECK_WITH_INFO(penaltyParams, "penaltyParams for setup is not set");
|
|
|
|
bool const useTemperature = mDecodingMode.isUseTemperature() && penaltyParams->temperature.has_value();
|
|
bool const useRepetitionPenalty
|
|
= mDecodingMode.isUseRepetitionPenalty() && penaltyParams->repetitionPenalty.has_value();
|
|
bool const usePresencePenalty = mDecodingMode.isUsePresencePenalty() && penaltyParams->presencePenalty.has_value();
|
|
bool const useFrequencyPenalty
|
|
= mDecodingMode.isUseFrequencyPenalty() && penaltyParams->frequencyPenalty.has_value();
|
|
bool const useMinLength = mDecodingMode.isUseMinLength() && penaltyParams->minLength.has_value();
|
|
// FIXME(nkorobov): once one of the requests has some penalty, we will always have to compute it.
|
|
// To avoid that we need to scan through all active requests at each iteration.
|
|
mUseTemperature |= useTemperature;
|
|
mUseRepetitionPenalty |= useRepetitionPenalty;
|
|
mUsePresencePenalty |= usePresencePenalty;
|
|
mUseFrequencyPenalty |= useFrequencyPenalty;
|
|
mUseMinLength |= useMinLength;
|
|
|
|
if (mUseTemperature)
|
|
{
|
|
fillBuffers(penaltyParams->temperature, DefaultDecodingParams::getTemperature(), mTemperature,
|
|
mTemperatureDevice, defaultBatchSlots, getLimitsPenalty(DecodingPenaltyType::Temperature),
|
|
"temperature penalty");
|
|
}
|
|
if (mUseRepetitionPenalty)
|
|
{
|
|
fillBuffers(penaltyParams->repetitionPenalty, DefaultDecodingParams::getRepetitionPenalty(), mRepetitionPenalty,
|
|
mRepetitionPenaltyDevice, defaultBatchSlots, getLimitsPenalty(DecodingPenaltyType::Repetition),
|
|
"repetition penalty");
|
|
}
|
|
if (mUsePresencePenalty)
|
|
{
|
|
fillBuffers(penaltyParams->presencePenalty, DefaultDecodingParams::getPresencePenalty(), mPresencePenalty,
|
|
mPresencePenaltyDevice, defaultBatchSlots, getLimitsPenalty(DecodingPenaltyType::Presence),
|
|
"presence penalty");
|
|
}
|
|
if (mUseFrequencyPenalty)
|
|
{
|
|
fillBuffers(penaltyParams->frequencyPenalty, DefaultDecodingParams::getFrequencyPenalty(), mFrequencyPenalty,
|
|
mFrequencyPenaltyDevice, defaultBatchSlots, getLimitsPenalty(DecodingPenaltyType::Frequency),
|
|
"frequency penalty");
|
|
}
|
|
if (mUseMinLength)
|
|
{
|
|
fillBuffers(penaltyParams->minLength, DefaultDecodingParams::getMinLength(), mMinLength, mMinLengthDevice,
|
|
defaultBatchSlots, getLimitsPenalty(DecodingPenaltyType::MinLength), "min length");
|
|
}
|
|
|
|
TLLM_LOG_TRACE("%s stop", __PRETTY_FUNCTION__);
|
|
}
|
|
|
|
template <typename T>
|
|
void PenaltyLayer<T>::forwardAsync(
|
|
std::shared_ptr<BaseDecodingOutputs> const& baseOutputs, std::shared_ptr<BaseDecodingInputs> const& baseInputs)
|
|
{
|
|
TLLM_LOG_TRACE("%s start", __PRETTY_FUNCTION__);
|
|
|
|
auto outputs = std::dynamic_pointer_cast<BaseDecodingOutputs>(baseOutputs);
|
|
auto params = std::dynamic_pointer_cast<DecodingInputs>(baseInputs);
|
|
|
|
auto const localDecoderDomain = getLocalDecoderDomain(params, mDecoderDomain);
|
|
auto const maxSeqLen = outputs->outputIds->getDimension<-1>();
|
|
auto batchSlots = bufferCastOrNull<SizeType32>(params->batchSlots);
|
|
|
|
if (!mLogitsPtrsHost->data())
|
|
{
|
|
mLogitsPtrsHost = mBufferManager->pinnedPool(
|
|
ITensor::makeShape({static_cast<int32_t>(maxSeqLen), static_cast<int32_t>(mDecoderDomain.getBatchSize())}),
|
|
TRTDataType<T*>::value);
|
|
mRuntimeMaxSeqLen = maxSeqLen;
|
|
}
|
|
|
|
mCyclicStep = mCyclicStep % mRuntimeMaxSeqLen;
|
|
|
|
TensorPtr logitsPtrsHost = ITensor::slice(mLogitsPtrsHost, mCyclicStep, 1);
|
|
auto logitsPtrsHostData = bufferCast<T*>(*logitsPtrsHost);
|
|
for (SizeType32 bi = 0; bi < localDecoderDomain.getBatchSize(); bi++)
|
|
{
|
|
if (params->logitsVec)
|
|
{
|
|
TLLM_CHECK_WITH_INFO(params->logitsVec->size() == localDecoderDomain.getBatchSize(),
|
|
"Logits vector size (%lu) is not equal to the batchSize (%d)", params->logitsVec->size(),
|
|
localDecoderDomain.getBatchSize());
|
|
logitsPtrsHostData[bi] = bufferCastOrNull<T>(params->logitsVec.value()[bi]);
|
|
}
|
|
else
|
|
{
|
|
TensorPtr logitsForBatchIndex = ITensor::slice(params->logits.value(), ITensor::makeShape({bi}));
|
|
auto const ptrToLogitsForBatchIndex = bufferCastOrNull<T>(logitsForBatchIndex);
|
|
logitsPtrsHostData[bi] = ptrToLogitsForBatchIndex;
|
|
}
|
|
}
|
|
|
|
auto inputLengths = bufferCastOrNull<SizeType32>(params->inputLengths);
|
|
auto embeddingBias = bufferCastOrNull<T>(params->embeddingBias);
|
|
auto batchSlotsHost = params->batchSlots ? params->batchSlots.value()
|
|
: getDefaultBatchSlots(localDecoderDomain.getBatchSize(), *mBufferManager);
|
|
auto batchSlotsHostPtr = bufferCast<SizeType32>(*batchSlotsHost);
|
|
#define GET_PENALTIES(capital_name, type) \
|
|
(mUse##capital_name \
|
|
&& !allOfBatchSlots(batchSlotsHostPtr, bufferCast<type>(*m##capital_name), localDecoderDomain.getBatchSize(), \
|
|
DefaultDecodingParams::get##capital_name())) \
|
|
? m##capital_name##Device \
|
|
: nullptr;
|
|
|
|
auto temperatures = GET_PENALTIES(Temperature, float);
|
|
auto repetitionPenalties = GET_PENALTIES(RepetitionPenalty, float);
|
|
auto presencePenalties = GET_PENALTIES(PresencePenalty, float);
|
|
auto frequencyPenalties = GET_PENALTIES(FrequencyPenalty, float);
|
|
auto minLengths = GET_PENALTIES(MinLength, SizeType32);
|
|
|
|
#undef GET_PENALTIES
|
|
|
|
auto const tokensPerStep = bufferCastOrNull<SizeType32>(params->curTokensPerStep);
|
|
|
|
InvokeBatchApplyPenaltyParams<T> penaltyParams;
|
|
penaltyParams.inputLogits = reinterpret_cast<T const* const*>(logitsPtrsHostData);
|
|
penaltyParams.outputLogits = bufferCast<T>(*mRuntimeLogitsDevice);
|
|
penaltyParams.biases = embeddingBias;
|
|
penaltyParams.penaltyWorkspace = bufferCastOrNull<TokenIdType>(mPenaltyWorkspaceDevice);
|
|
penaltyParams.penaltyWorkspacePrev = bufferCastOrNull<TokenIdType>(mPenaltyWorkspacePrevDevice);
|
|
penaltyParams.temperatures = bufferCastOrNull<float>(temperatures);
|
|
penaltyParams.repetitionPenalties = bufferCastOrNull<float>(repetitionPenalties);
|
|
penaltyParams.presencePenalties = bufferCastOrNull<float>(presencePenalties);
|
|
penaltyParams.frequencyPenalties = bufferCastOrNull<float>(frequencyPenalties);
|
|
penaltyParams.batchSize = localDecoderDomain.getBatchSize();
|
|
penaltyParams.beamWidth = localDecoderDomain.getBeamWidth();
|
|
penaltyParams.maxSeqLen = maxSeqLen;
|
|
penaltyParams.vocabSize = mDecoderDomain.getVocabSize();
|
|
penaltyParams.vocabSizePadded = mDecoderDomain.getVocabSizePadded();
|
|
penaltyParams.outputIdsPtr = bufferCast<TokenIdType const*>(*outputs->outputIdsPtr);
|
|
penaltyParams.parentIdsPtr = bufferCast<SizeType32 const*>(*outputs->parentIdsPtr);
|
|
penaltyParams.inputLengths = inputLengths;
|
|
penaltyParams.sequenceLengths = bufferCast<SizeType32>(*outputs->sequenceLength.value());
|
|
penaltyParams.minLengths = bufferCastOrNull<SizeType32>(minLengths);
|
|
penaltyParams.endIds = bufferCast<TokenIdType>(*params->endIds);
|
|
penaltyParams.batchSlots = batchSlots;
|
|
penaltyParams.maxTokensPerStep = mDecoderDomain.getMaxDecodingTokens();
|
|
penaltyParams.tokensPerStep = tokensPerStep;
|
|
penaltyParams.stream = getStream();
|
|
invokeBatchApplyPenalty(penaltyParams);
|
|
sync_check_cuda_error();
|
|
|
|
mCyclicStep += 1;
|
|
|
|
auto const logitsShape = ITensor::makeShape({localDecoderDomain.getBatchSize(),
|
|
mDecoderDomain.getMaxDecodingTokens(), localDecoderDomain.getBeamWidth(), mDecoderDomain.getVocabSizePadded()});
|
|
params->logits = ITensor::view(mRuntimeLogitsDevice, logitsShape);
|
|
|
|
if (mDecodingMode.isBeamSearch())
|
|
{
|
|
std::swap(mPenaltyWorkspaceDevice, mPenaltyWorkspacePrevDevice);
|
|
}
|
|
|
|
TLLM_LOG_TRACE("%s stop", __PRETTY_FUNCTION__);
|
|
}
|
|
|
|
template class PenaltyLayer<float>;
|
|
template class PenaltyLayer<half>;
|
|
|
|
} // namespace tensorrt_llm::layers
|