experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Backward compatible api change:BoostedTreesUpdateEnsembleV2 works on list of feature_ids.

Browse Source 
PiperOrigin-RevId: 289687663
Change-Id: I5d12d044ae42fc34f03a3eaa357bf71b7cb06eec
This commit is contained in:
A. Unique TensorFlower 2020-01-14 11:02:55 -08:00 committed by TensorFlower Gardener
parent 857b55bc2d
commit 8a2a86318b
4 changed files with 35 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
tensorflow/core/api_def/base_api/api_def_BoostedTreesUpdateEnsembleV2.pbtxt
View File

@ -91,6 +91,14 @@ END
name: "logits_dimension" name: "logits_dimension"
description: <<END description: <<END
scalar, dimension of the logits scalar, dimension of the logits
END
}
attr {
name: "num_groups"
description: <<END
Number of groups of split information to process, where a group contains feature
ids that are processed together in BoostedTreesCalculateBestFeatureSplitOpV2.
INFERRED.
END END
} }
summary: "Updates the tree ensemble by adding a layer to the last tree being grown" summary: "Updates the tree ensemble by adding a layer to the last tree being grown"

8
tensorflow/core/kernels/boosted_trees/training_ops.cc
View File

@ -269,9 +269,11 @@ class BoostedTreesUpdateEnsembleV2Op : public OpKernel {
OP_REQUIRES_OK(context, OP_REQUIRES_OK(context,
context->input_list("split_types", &split_types_list)); context->input_list("split_types", &split_types_list));
const Tensor* feature_ids_t; OpInputList feature_ids_list;
OP_REQUIRES_OK(context, context->input("feature_ids", &feature_ids_t)); OP_REQUIRES_OK(context,
const auto feature_ids = feature_ids_t->vec<int32>(); context->input_list("feature_ids", &feature_ids_list));
// TODO(crawles): Read groups of feature ids and find best splits among all.
const auto feature_ids = feature_ids_list[0].vec<int32>();
const Tensor* max_depth_t; const Tensor* max_depth_t;
OP_REQUIRES_OK(context, context->input("max_depth", &max_depth_t)); OP_REQUIRES_OK(context, context->input("max_depth", &max_depth_t));

7
tensorflow/core/ops/boosted_trees_ops.cc
View File

@ -618,7 +618,7 @@ REGISTER_OP("BoostedTreesUpdateEnsemble")
REGISTER_OP("BoostedTreesUpdateEnsembleV2") REGISTER_OP("BoostedTreesUpdateEnsembleV2")
.Input("tree_ensemble_handle: resource") .Input("tree_ensemble_handle: resource")
.Input("feature_ids: int32") .Input("feature_ids: num_groups * int32")
.Input("dimension_ids: num_features * int32") .Input("dimension_ids: num_features * int32")
.Input("node_ids: num_features * int32") .Input("node_ids: num_features * int32")
.Input("gains: num_features * float") .Input("gains: num_features * float")
@ -631,13 +631,18 @@ REGISTER_OP("BoostedTreesUpdateEnsembleV2")
.Input("pruning_mode: int32") .Input("pruning_mode: int32")
.Attr("num_features: int >= 0") // Inferred. .Attr("num_features: int >= 0") // Inferred.
.Attr("logits_dimension: int = 1") .Attr("logits_dimension: int = 1")
.Attr("num_groups: int = 1") // Number of groups to process.
.SetShapeFn([](shape_inference::InferenceContext* c) { .SetShapeFn([](shape_inference::InferenceContext* c) {
shape_inference::ShapeHandle shape_handle; shape_inference::ShapeHandle shape_handle;
int num_features; int num_features;
TF_RETURN_IF_ERROR(c->GetAttr("num_features", &num_features)); TF_RETURN_IF_ERROR(c->GetAttr("num_features", &num_features));
int num_groups;
TF_RETURN_IF_ERROR(c->GetAttr("num_groups", &num_groups));
// Feature_ids, should be one for each feature. // Feature_ids, should be one for each feature.
shape_inference::ShapeHandle feature_ids_shape; shape_inference::ShapeHandle feature_ids_shape;
// TODO(crawles): remove 1 hardcode once kernel operates on multiple
// groups.
TF_RETURN_IF_ERROR(c->WithRank(c->input(1), 1, &feature_ids_shape)); TF_RETURN_IF_ERROR(c->WithRank(c->input(1), 1, &feature_ids_shape));
TF_RETURN_IF_ERROR( TF_RETURN_IF_ERROR(
c->Merge(c->input(1), c->Vector(num_features), &shape_handle)); c->Merge(c->input(1), c->Vector(num_features), &shape_handle));

32
tensorflow/python/kernel_tests/boosted_trees/training_ops_test.py
View File

@ -180,7 +180,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
# Tree will be finalized now, since we will reach depth 1. # Tree will be finalized now, since we will reach depth 1.
max_depth=1, max_depth=1,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions, feature2_dimensions], dimension_ids=[feature1_dimensions, feature2_dimensions],
node_ids=[feature1_nodes, feature2_nodes], node_ids=[feature1_nodes, feature2_nodes],
gains=[feature1_gains, feature2_gains], gains=[feature1_gains, feature2_gains],
@ -289,7 +289,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
# Tree will be finalized now, since we will reach depth 1. # Tree will be finalized now, since we will reach depth 1.
max_depth=1, max_depth=1,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions, feature2_dimensions], dimension_ids=[feature1_dimensions, feature2_dimensions],
node_ids=[feature1_nodes, feature2_nodes], node_ids=[feature1_nodes, feature2_nodes],
gains=[feature1_gains, feature2_gains], gains=[feature1_gains, feature2_gains],
@ -401,7 +401,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
# Tree will be finalized now, since we will reach depth 1. # Tree will be finalized now, since we will reach depth 1.
max_depth=1, max_depth=1,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions, feature2_dimensions], dimension_ids=[feature1_dimensions, feature2_dimensions],
node_ids=[feature1_nodes, feature2_nodes], node_ids=[feature1_nodes, feature2_nodes],
gains=[feature1_gains, feature2_gains], gains=[feature1_gains, feature2_gains],
@ -809,7 +809,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
# tree is going to be finalized now, since we reach depth 2. # tree is going to be finalized now, since we reach depth 2.
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[ dimension_ids=[
feature1_dimensions, feature2_dimensions, feature3_dimensions feature1_dimensions, feature2_dimensions, feature3_dimensions
], ],
@ -1014,7 +1014,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
# tree is going to be finalized now, since we reach depth 2. # tree is going to be finalized now, since we reach depth 2.
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[ dimension_ids=[
feature1_dimensions, feature2_dimensions, feature3_dimensions feature1_dimensions, feature2_dimensions, feature3_dimensions
], ],
@ -1230,7 +1230,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
# tree is going to be finalized now, since we reach depth 2. # tree is going to be finalized now, since we reach depth 2.
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[ dimension_ids=[
feature1_dimensions, feature2_dimensions, feature3_dimensions feature1_dimensions, feature2_dimensions, feature3_dimensions
], ],
@ -1610,7 +1610,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
learning_rate=0.1, learning_rate=0.1,
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions], dimension_ids=[feature1_dimensions],
node_ids=[feature1_nodes], node_ids=[feature1_nodes],
gains=[feature1_gains], gains=[feature1_gains],
@ -1769,7 +1769,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
learning_rate=0.1, learning_rate=0.1,
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions], dimension_ids=[feature1_dimensions],
node_ids=[feature1_nodes], node_ids=[feature1_nodes],
gains=[feature1_gains], gains=[feature1_gains],
@ -1942,7 +1942,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.NO_PRUNING,
learning_rate=0.1, learning_rate=0.1,
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions], dimension_ids=[feature1_dimensions],
node_ids=[feature1_nodes], node_ids=[feature1_nodes],
gains=[feature1_gains], gains=[feature1_gains],
@ -2309,7 +2309,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
pruning_mode=boosted_trees_ops.PruningMode.PRE_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.PRE_PRUNING,
# tree is going to be finalized now, since we reach depth 2. # tree is going to be finalized now, since we reach depth 2.
max_depth=3, max_depth=3,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[ dimension_ids=[
feature1_dimensions, feature2_dimensions, feature3_dimensions feature1_dimensions, feature2_dimensions, feature3_dimensions
], ],
@ -3041,7 +3041,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
learning_rate=1.0, learning_rate=1.0,
pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING,
max_depth=3, max_depth=3,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions, feature2_dimensions], dimension_ids=[feature1_dimensions, feature2_dimensions],
node_ids=[feature1_nodes, feature2_nodes], node_ids=[feature1_nodes, feature2_nodes],
gains=[feature1_gains, feature2_gains], gains=[feature1_gains, feature2_gains],
@ -3140,7 +3140,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
learning_rate=1.0, learning_rate=1.0,
pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING,
max_depth=3, max_depth=3,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions], dimension_ids=[feature1_dimensions],
node_ids=[feature1_nodes], node_ids=[feature1_nodes],
gains=[feature1_gains], gains=[feature1_gains],
@ -3293,7 +3293,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
learning_rate=1.0, learning_rate=1.0,
pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING,
max_depth=3, max_depth=3,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions], dimension_ids=[feature1_dimensions],
node_ids=[feature1_nodes], node_ids=[feature1_nodes],
gains=[feature1_gains], gains=[feature1_gains],
@ -3679,7 +3679,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
learning_rate=1.0, learning_rate=1.0,
pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING,
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions, feature2_dimensions], dimension_ids=[feature1_dimensions, feature2_dimensions],
node_ids=[feature1_nodes, feature2_nodes], node_ids=[feature1_nodes, feature2_nodes],
gains=[feature1_gains, feature2_gains], gains=[feature1_gains, feature2_gains],
@ -3778,7 +3778,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
learning_rate=1.0, learning_rate=1.0,
pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING,
max_depth=2, max_depth=2,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions], dimension_ids=[feature1_dimensions],
node_ids=[feature1_nodes], node_ids=[feature1_nodes],
gains=[feature1_gains], gains=[feature1_gains],
@ -4014,7 +4014,7 @@ class UpdateTreeEnsembleOpTest(test_util.TensorFlowTestCase):
learning_rate=1.0, learning_rate=1.0,
pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING, pruning_mode=boosted_trees_ops.PruningMode.POST_PRUNING,
max_depth=1, max_depth=1,
feature_ids=feature_ids, feature_ids=[feature_ids],
dimension_ids=[feature1_dimensions, feature2_dimensions], dimension_ids=[feature1_dimensions, feature2_dimensions],
node_ids=[feature1_nodes, feature2_nodes], node_ids=[feature1_nodes, feature2_nodes],
gains=[feature1_gains, feature2_gains], gains=[feature1_gains, feature2_gains],