void TCatboostOptions::Validate() const {
SystemOptions.Get().Validate();
BoostingOptions.Get().Validate();
ObliviousTreeOptions.Get().Validate();
ELossFunction lossFunction = LossFunctionDescription->GetLossFunction();
{
const ui32 classesCount = DataProcessingOptions->ClassesCount;
if (classesCount != 0 ) {
CB_ENSURE(IsMultiClassError(lossFunction), "classes_count parameter takes effect only with MultiClass/MultiClassOneVsAll loss functions");
CB_ENSURE(classesCount > 1, "classes-count should be at least 2");
}
const auto& classWeights = DataProcessingOptions->ClassWeights.Get();
if (!classWeights.empty()) {
CB_ENSURE(lossFunction == ELossFunction::Logloss || IsMultiClassError(lossFunction),
"class weights takes effect only with Logloss, MultiClass and MultiClassOneVsAll loss functions");
CB_ENSURE(IsMultiClassError(lossFunction) || (classWeights.size() == 2),
"if loss-function is Logloss, then class weights should be given for 0 and 1 classes");
CB_ENSURE(classesCount == 0 || classesCount == classWeights.size(), "class weights should be specified for each class in range 0, ... , classes_count - 1");
}
}
ELeavesEstimation leavesEstimation = ObliviousTreeOptions->LeavesEstimationMethod;
if (lossFunction == ELossFunction::Quantile ||
lossFunction == ELossFunction::MAE ||
lossFunction == ELossFunction::LogLinQuantile ||
lossFunction == ELossFunction::MAPE)
{
CB_ENSURE(leavesEstimation != ELeavesEstimation::Newton,
"Newton leave estimation method is not supported for " << lossFunction << " loss function");
CB_ENSURE(ObliviousTreeOptions->LeavesEstimationIterations == 1U,
"gradient_iterations should equals 1 for this mode");
}
if (GetTaskType() == ETaskType::CPU) {
CB_ENSURE(!(IsQuerywiseError(lossFunction) && leavesEstimation == ELeavesEstimation::Newton),
"This leaf estimation method is not supported for querywise error for CPU learning");
CB_ENSURE(!(IsPairwiseError(lossFunction) && leavesEstimation == ELeavesEstimation::Newton),
"This leaf estimation method is not supported for pairwise error");
}
ValidateCtrs(CatFeatureParams->SimpleCtrs, lossFunction, false);
for (const auto& perFeatureCtr : CatFeatureParams->PerFeatureCtrs.Get()) {
ValidateCtrs(perFeatureCtr.second, lossFunction, false);
}
ValidateCtrs(CatFeatureParams->CombinationCtrs, lossFunction, true);
}
void TLearnContext::InitContext(const TDataset& learnData, const TDatasetPtrs& testDataPtrs) {
LearnProgress.PoolCheckSum = CalcFeaturesCheckSum(learnData.AllFeatures);
for (const TDataset* testData : testDataPtrs) {
LearnProgress.PoolCheckSum += CalcFeaturesCheckSum(testData->AllFeatures);
}
auto lossFunction = Params.LossFunctionDescription->GetLossFunction();
int foldCount = Max<ui32>(Params.BoostingOptions->PermutationCount - 1, 1);
const bool noCtrs = IsCategoricalFeaturesEmpty(learnData.AllFeatures);
if (Params.BoostingOptions->BoostingType == EBoostingType::Plain && noCtrs) {
foldCount = 1;
}
LearnProgress.Folds.reserve(foldCount);
UpdateCtrsTargetBordersOption(lossFunction, LearnProgress.ApproxDimension, &Params.CatFeatureParams.Get());
CtrsHelper.InitCtrHelper(Params.CatFeatureParams,
Layout,
learnData.Target,
lossFunction,
ObjectiveDescriptor);
//Todo(noxoomo): check and init
const auto& boostingOptions = Params.BoostingOptions.Get();
ui32 foldPermutationBlockSize = boostingOptions.PermutationBlockSize;
if (foldPermutationBlockSize == FoldPermutationBlockSizeNotSet) {
foldPermutationBlockSize = DefaultFoldPermutationBlockSize(learnData.GetSampleCount());
}
if (IsPlainMode(Params.BoostingOptions->BoostingType) && noCtrs) {
foldPermutationBlockSize = learnData.GetSampleCount();
}
const auto storeExpApproxes = IsStoreExpApprox(Params.LossFunctionDescription->GetLossFunction());
const bool hasPairwiseWeights = IsPairwiseError(Params.LossFunctionDescription->GetLossFunction());
if (IsPlainMode(Params.BoostingOptions->BoostingType)) {
for (int foldIdx = 0; foldIdx < foldCount; ++foldIdx) {
LearnProgress.Folds.emplace_back(
BuildPlainFold(
learnData,
CtrsHelper.GetTargetClassifiers(),
foldIdx != 0,
(Params.SystemOptions->IsSingleHost() ? foldPermutationBlockSize : learnData.GetSampleCount()),
LearnProgress.ApproxDimension,
storeExpApproxes,
hasPairwiseWeights,
Rand
)
);
}
} else {
for (int foldIdx = 0; foldIdx < foldCount; ++foldIdx) {
LearnProgress.Folds.emplace_back(
BuildDynamicFold(
learnData,
CtrsHelper.GetTargetClassifiers(),
foldIdx != 0,
foldPermutationBlockSize,
LearnProgress.ApproxDimension,
boostingOptions.FoldLenMultiplier,
storeExpApproxes,
hasPairwiseWeights,
Rand
)
);
}
}
LearnProgress.AveragingFold = BuildPlainFold(
learnData,
CtrsHelper.GetTargetClassifiers(),
!(Params.DataProcessingOptions->HasTimeFlag),
/*permuteBlockSize=*/ (Params.SystemOptions->IsSingleHost() ? foldPermutationBlockSize : learnData.GetSampleCount()),
LearnProgress.ApproxDimension,
storeExpApproxes,
hasPairwiseWeights,
Rand
);
LearnProgress.AvrgApprox.resize(LearnProgress.ApproxDimension, TVector<double>(learnData.GetSampleCount()));
if (!learnData.Baseline.empty()) {
LearnProgress.AvrgApprox = learnData.Baseline;
}
ResizeRank2(testDataPtrs.size(), LearnProgress.ApproxDimension, LearnProgress.TestApprox);
for (size_t testIdx = 0; testIdx < testDataPtrs.size(); ++testIdx) {
const auto* testData = testDataPtrs[testIdx];
if (testData == nullptr || testData->GetSampleCount() == 0) {
continue;
}
if (testData->Baseline.empty()) {
for (auto& approxDim : LearnProgress.TestApprox[testIdx]) {
approxDim.resize(testData->GetSampleCount());
}
} else {
LearnProgress.TestApprox[testIdx] = testData->Baseline;
}
}
}