Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Test categorical features with column-split gpu quantile (#9595)

Browse Source
This commit is contained in:
Rong Ou
2023-09-22 18:55:09 -07:00
committed by GitHub
parent a90d204942
commit def77870f3
3 changed files with 70 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
tests/cpp/common/test_quantile.cu
View File

@@ -339,6 +339,31 @@ TEST(GPUQuantile, MultiMerge) {
});
}
TEST(GPUQuantile, MissingColumns) {
auto dmat = std::unique_ptr<DMatrix>{[=]() {
std::size_t constexpr kRows = 1000, kCols = 100;
auto sparsity = 0.5f;
std::vector<FeatureType> ft(kCols);
for (size_t i = 0; i < ft.size(); ++i) {
ft[i] = (i % 2 == 0) ? FeatureType::kNumerical : FeatureType::kCategorical;
}
auto dmat = RandomDataGenerator{kRows, kCols, sparsity}
.Seed(0)
.Lower(.0f)
.Upper(1.0f)
.Type(ft)
.MaxCategory(13)
.GenerateDMatrix();
return dmat->SliceCol(2, 1);
}()};
dmat->Info().data_split_mode = DataSplitMode::kRow;
auto ctx = MakeCUDACtx(0);
std::size_t constexpr kBins = 64;
HistogramCuts cuts = common::DeviceSketch(&ctx, dmat.get(), kBins);
ASSERT_TRUE(cuts.HasCategorical());
}
namespace {
void TestAllReduceBasic() {
auto const world = collective::GetWorldSize();
@@ -422,18 +447,14 @@ TEST_F(MGPUQuantileTest, AllReduceBasic) {
}
namespace {
void TestColumnSplitBasic() {
void TestColumnSplit(DMatrix* dmat) {
auto const world = collective::GetWorldSize();
auto const rank = collective::GetRank();
std::size_t constexpr kRows = 1000, kCols = 100, kBins = 64;
auto m = std::unique_ptr<DMatrix>{[=]() {
auto dmat = RandomDataGenerator{kRows, kCols, 0}.GenerateDMatrix();
return dmat->SliceCol(world, rank);
}()};
auto m = std::unique_ptr<DMatrix>{dmat->SliceCol(world, rank)};
// Generate cuts for distributed environment.
auto ctx = MakeCUDACtx(GPUIDX);
std::size_t constexpr kBins = 64;
HistogramCuts distributed_cuts = common::DeviceSketch(&ctx, m.get(), kBins);
// Generate cuts for single node environment
@@ -466,7 +487,26 @@ void TestColumnSplitBasic() {
} // anonymous namespace
TEST_F(MGPUQuantileTest, ColumnSplitBasic) {
DoTest(TestColumnSplitBasic);
std::size_t constexpr kRows = 1000, kCols = 100;
auto dmat = RandomDataGenerator{kRows, kCols, 0}.GenerateDMatrix();
DoTest(TestColumnSplit, dmat.get());
}
TEST_F(MGPUQuantileTest, ColumnSplitCategorical) {
std::size_t constexpr kRows = 1000, kCols = 100;
auto sparsity = 0.5f;
std::vector<FeatureType> ft(kCols);
for (size_t i = 0; i < ft.size(); ++i) {
ft[i] = (i % 2 == 0) ? FeatureType::kNumerical : FeatureType::kCategorical;
}
auto dmat = RandomDataGenerator{kRows, kCols, sparsity}
.Seed(0)
.Lower(.0f)
.Upper(1.0f)
.Type(ft)
.MaxCategory(13)
.GenerateDMatrix();
DoTest(TestColumnSplit, dmat.get());
}
namespace {