17697974118

Committed 13 Sep 2025 02:36PM UTC coverage: 60.51% (+1.2%) from 59.335%

Build # 17697974118

Build Type

Pull #219

github

Committed by

web-flow

Commit Message

Merge a1c5ecbc6 into 6c1992b40

Pull Request Pull Request #219: stdlib Library Nodes and ConstantNodes

Run Details

565 of 1799 new or added lines in 102 files covered. (31.41%)

102 existing lines in 38 files now uncovered.

9442 of 15604 relevant lines covered (60.51%)

107.02 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

0.0

/src/data_flow/library_nodes/math/ml/batch_normalization.cpp

#include "sdfg/data_flow/library_nodes/math/ml/batch_normalization.h"

#include "sdfg/analysis/analysis.h"
#include "sdfg/analysis/scope_analysis.h"
#include "sdfg/builder/structured_sdfg_builder.h"

namespace sdfg {
namespace math {
namespace ml {

BatchNormalizationNode::BatchNormalizationNode(
    size_t element_id,
    const DebugInfo &debug_info,
    const graph::Vertex vertex,
    data_flow::DataFlowGraph &parent,
    const std::vector<symbolic::Expression> &shape,
    int axis,
    const std::string &epsilon
)
    : MathNode(
          element_id,
          debug_info,
          vertex,
          parent,
          LibraryNodeType_BatchNormalization,
          {"Y"},
          {"X", "Scale", "B", "input_mean", "input_var"},
          data_flow::ImplementationType_NONE
      ),
      shape_(shape), axis_(axis), epsilon_(epsilon) {}

symbolic::SymbolSet BatchNormalizationNode::symbols() const {
    symbolic::SymbolSet syms;
    for (const auto &dim : shape_) {
        for (auto &atom : symbolic::atoms(dim)) {
            syms.insert(atom);
        }
    }
    return syms;
}

void BatchNormalizationNode::replace(const symbolic::Expression old_expression, const symbolic::Expression new_expression) {
    for (auto &dim : shape_) {
        dim = symbolic::subs(dim, old_expression, new_expression);
    }
}

void BatchNormalizationNode::validate(const Function &) const {}

bool BatchNormalizationNode::expand(builder::StructuredSDFGBuilder &builder, analysis::AnalysisManager &analysis_manager) {
    auto &dataflow = this->get_parent();
    auto &block = static_cast<structured_control_flow::Block &>(*dataflow.get_parent());

    auto &scope_analysis = analysis_manager.get<analysis::ScopeAnalysis>();
    auto &parent = static_cast<structured_control_flow::Sequence &>(*scope_analysis.parent_scope(&block));
    int index = parent.index(block);
    auto &transition = parent.at(index).second;

    // Locate edges
    const data_flow::Memlet *iedge_input = nullptr;
    const data_flow::Memlet *iedge_scale = nullptr;
    const data_flow::Memlet *iedge_bias = nullptr;
    const data_flow::Memlet *iedge_mean = nullptr;
    const data_flow::Memlet *iedge_var = nullptr;
    const data_flow::Memlet *oedge_output = nullptr;
    for (const auto &edge : dataflow.in_edges(*this)) {
        if (edge.dst_conn() == "X") {
            iedge_input = &edge;
        } else if (edge.dst_conn() == "Scale") {
            iedge_scale = &edge;
        } else if (edge.dst_conn() == "B") {
            iedge_bias = &edge;
        } else if (edge.dst_conn() == "input_mean") {
            iedge_mean = &edge;
        } else if (edge.dst_conn() == "input_var") {
            iedge_var = &edge;
        }
    }
    for (const auto &edge : dataflow.out_edges(*this)) {
        if (edge.src_conn() == "Y") {
            oedge_output = &edge;
        }
    }
    if (!iedge_input || !iedge_scale || !iedge_bias || !iedge_mean || !iedge_var || !oedge_output) return false;

    std::string input_name = static_cast<const data_flow::AccessNode &>(iedge_input->src()).data();
    std::string scale_name = static_cast<const data_flow::AccessNode &>(iedge_scale->src()).data();
    std::string bias_name = static_cast<const data_flow::AccessNode &>(iedge_bias->src()).data();
    std::string mean_name = static_cast<const data_flow::AccessNode &>(iedge_mean->src()).data();
    std::string var_name = static_cast<const data_flow::AccessNode &>(iedge_var->src()).data();
    std::string output_name = static_cast<const data_flow::AccessNode &>(oedge_output->dst()).data();

    // Create new sequence before
    auto &new_sequence = builder.add_sequence_before(parent, block, transition.assignments(), block.debug_info());
    structured_control_flow::Sequence *last_scope = &new_sequence;

    std::vector<symbolic::Expression> loop_syms;
    structured_control_flow::Map *last_map = nullptr;
    for (size_t d = 0; d < this->shape_.size(); ++d) {
        std::string indvar_str = builder.find_new_name("_i");
        builder.add_container(indvar_str, types::Scalar(types::PrimitiveType::UInt64));
        auto indvar = symbolic::symbol(indvar_str);
        auto init = symbolic::zero();
        auto update = symbolic::add(indvar, symbolic::one());
        auto cond = symbolic::Lt(indvar, this->shape_[d]);
        last_map = &builder.add_map(
            *last_scope,
            indvar,
            cond,
            init,
            update,
            structured_control_flow::ScheduleType_Sequential::create(),
            {},
            block.debug_info()
        );
        last_scope = &last_map->root();
        loop_syms.push_back(indvar);
    }

    // Create normalization block
    auto &norm_block = builder.add_block(*last_scope);

    // Create access nodes for normalization
    auto &input_access_norm = builder.add_access(norm_block, input_name);
    auto &scale_access_norm = builder.add_access(norm_block, scale_name);
    auto &bias_access_norm = builder.add_access(norm_block, bias_name);
    auto &mean_access_norm = builder.add_access(norm_block, mean_name);
    auto &var_access_norm = builder.add_access(norm_block, var_name);
    auto &output_access_norm = builder.add_access(norm_block, output_name);

    // Add epsilon to variance and compute standard deviation
    auto &add_epsilon_tasklet =
        builder.add_tasklet(norm_block, data_flow::TaskletCode::add, "_out", {"_in1", epsilon_});
    auto &var_eps_access = builder.add_access(norm_block, builder.find_new_name("_var_eps"));
    builder.add_computational_memlet(
        norm_block, var_access_norm, add_epsilon_tasklet, "_in1", loop_syms, iedge_var->base_type()
    );
    builder
        .add_computational_memlet(norm_block, add_epsilon_tasklet, "_out", var_eps_access, {}, iedge_var->base_type());

    auto &sqrt_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::sqrt, "_out", {"_in"});
    auto &std_dev_access = builder.add_access(norm_block, builder.find_new_name("_std_dev"));
    builder.add_computational_memlet(norm_block, var_eps_access, sqrt_tasklet, "_in", {}, iedge_var->base_type());
    builder.add_computational_memlet(norm_block, sqrt_tasklet, "_out", std_dev_access, {}, iedge_var->base_type());

    // Normalize: (x - mean) / std_dev
    auto &sub_norm_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::sub, "_out", {"_in1", "_in2"});
    auto &centered_access = builder.add_access(norm_block, builder.find_new_name("_centered"));
    builder.add_computational_memlet(
        norm_block, input_access_norm, sub_norm_tasklet, "_in1", loop_syms, iedge_input->base_type()
    );
    builder.add_computational_memlet(
        norm_block, mean_access_norm, sub_norm_tasklet, "_in2", loop_syms, iedge_mean->base_type()
    );
    builder
        .add_computational_memlet(norm_block, sub_norm_tasklet, "_out", centered_access, {}, iedge_input->base_type());

    auto &div_norm_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::div, "_out", {"_in1", "_in2"});
    auto &normalized_access = builder.add_access(norm_block, builder.find_new_name("_normalized"));
    builder
        .add_computational_memlet(norm_block, centered_access, div_norm_tasklet, "_in1", {}, iedge_input->base_type());
    builder
        .add_computational_memlet(norm_block, std_dev_access, div_norm_tasklet, "_in2", loop_syms, iedge_var->base_type());
    builder
        .add_computational_memlet(norm_block, div_norm_tasklet, "_out", normalized_access, {}, iedge_input->base_type());

    // Apply scale and bias: scale * normalized + bias
    auto &mul_scale_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::mul, "_out", {"_in1", "_in2"});
    auto &scaled_access = builder.add_access(norm_block, builder.find_new_name("_scaled"));
    builder
        .add_computational_memlet(norm_block, normalized_access, mul_scale_tasklet, "_in1", {}, iedge_input->base_type());
    builder.add_computational_memlet(
        norm_block, scale_access_norm, mul_scale_tasklet, "_in2", loop_syms, iedge_scale->base_type()
    );
    builder.add_computational_memlet(norm_block, mul_scale_tasklet, "_out", scaled_access, {}, iedge_input->base_type());

    auto &add_bias_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::add, "_out", {"_in1", "_in2"});
    builder.add_computational_memlet(norm_block, scaled_access, add_bias_tasklet, "_in1", {}, iedge_input->base_type());
    builder.add_computational_memlet(
        norm_block, bias_access_norm, add_bias_tasklet, "_in2", loop_syms, iedge_bias->base_type()
    );
    builder.add_computational_memlet(
        norm_block, add_bias_tasklet, "_out", output_access_norm, loop_syms, oedge_output->base_type()
    );

    // Cleanup old block
    builder.remove_memlet(block, *iedge_input);
    builder.remove_memlet(block, *iedge_scale);
    if (iedge_bias) {
        builder.remove_memlet(block, *iedge_bias);
    }
    builder.remove_memlet(block, *iedge_mean);
    builder.remove_memlet(block, *iedge_var);
    builder.remove_memlet(block, *oedge_output);
    builder.remove_node(block, *this);
    builder.remove_child(parent, index + 1);

    return true;
}

std::unique_ptr<data_flow::DataFlowNode> BatchNormalizationNode::
    clone(size_t element_id, const graph::Vertex vertex, data_flow::DataFlowGraph &parent) const {
    return std::unique_ptr<data_flow::DataFlowNode>(
        new BatchNormalizationNode(element_id, this->debug_info(), vertex, parent, this->shape_, axis_, epsilon_)
    );
}

nlohmann::json BatchNormalizationNodeSerializer::serialize(const data_flow::LibraryNode &library_node) {
    const BatchNormalizationNode &node = static_cast<const BatchNormalizationNode &>(library_node);
    nlohmann::json j;

    j["code"] = node.code().value();
    j["axis"] = node.axis();
    j["epsilon"] = node.epsilon();

    serializer::JSONSerializer serializer;
    j["shape"] = nlohmann::json::array();
    for (auto &dim : node.shape()) {
        j["shape"].push_back(serializer.expression(dim));
    }

    return j;
}

data_flow::LibraryNode &BatchNormalizationNodeSerializer::deserialize(
    const nlohmann::json &j, builder::StructuredSDFGBuilder &builder, structured_control_flow::Block &parent
) {
    auto code = j["code"].get<std::string>();
    if (code != LibraryNodeType_BatchNormalization.value()) {
        throw std::runtime_error("Invalid library node code");
    }

    sdfg::serializer::JSONSerializer serializer;
    DebugInfo debug_info = serializer.json_to_debug_info(j["debug_info"]);

    std::vector<symbolic::Expression> shape;
    for (const auto &dim : j["shape"]) {
        shape.push_back(symbolic::parse(dim.get<std::string>()));
    }

    auto axis = j["axis"].get<int>();
    auto epsilon = j["epsilon"].get<std::string>();

    return builder.add_library_node<BatchNormalizationNode>(parent, debug_info, shape, axis, epsilon);
}

} // namespace ml
} // namespace math
} // namespace sdfg

1	#include "sdfg/data_flow/library_nodes/math/ml/batch_normalization.h"
2
3	#include "sdfg/analysis/analysis.h"
4	#include "sdfg/analysis/scope_analysis.h"
5	#include "sdfg/builder/structured_sdfg_builder.h"
6
7	namespace sdfg {
8	namespace math {
9	namespace ml {
10
11	BatchNormalizationNode::BatchNormalizationNode(	×
12	size_t element_id,
13	const DebugInfo &debug_info,
14	const graph::Vertex vertex,
15	data_flow::DataFlowGraph &parent,
16	const std::vector<symbolic::Expression> &shape,
17	int axis,
18	const std::string &epsilon
19	)
20	: MathNode(	×
21	element_id,	×
22	debug_info,	×
23	vertex,	×
24	parent,	×
25	LibraryNodeType_BatchNormalization,
26	{"Y"},	×
27	{"X", "Scale", "B", "input_mean", "input_var"},	×
28	data_flow::ImplementationType_NONE
29	),
NEW 30	shape_(shape), axis_(axis), epsilon_(epsilon) {}	×
31
NEW 32	symbolic::SymbolSet BatchNormalizationNode::symbols() const {	×
NEW 33	symbolic::SymbolSet syms;	×
NEW 34	for (const auto &dim : shape_) {	×
NEW 35	for (auto &atom : symbolic::atoms(dim)) {	×
NEW 36	syms.insert(atom);	×
37	}
38	}
NEW 39	return syms;	×
NEW 40	}	×
41
NEW 42	void BatchNormalizationNode::replace(const symbolic::Expression old_expression, const symbolic::Expression new_expression) {	×
NEW 43	for (auto &dim : shape_) {	×
NEW 44	dim = symbolic::subs(dim, old_expression, new_expression);	×
45	}
NEW 46	}	×
47
NEW 48	void BatchNormalizationNode::validate(const Function &) const {}	×
49
50	bool BatchNormalizationNode::expand(builder::StructuredSDFGBuilder &builder, analysis::AnalysisManager &analysis_manager) {	×
51	auto &dataflow = this->get_parent();	×
52	auto &block = static_cast<structured_control_flow::Block &>(*dataflow.get_parent());	×
53
54	auto &scope_analysis = analysis_manager.get<analysis::ScopeAnalysis>();	×
55	auto &parent = static_cast<structured_control_flow::Sequence &>(*scope_analysis.parent_scope(&block));	×
56	int index = parent.index(block);	×
57	auto &transition = parent.at(index).second;	×
58
59	// Locate edges
60	const data_flow::Memlet *iedge_input = nullptr;	×
61	const data_flow::Memlet *iedge_scale = nullptr;	×
62	const data_flow::Memlet *iedge_bias = nullptr;	×
63	const data_flow::Memlet *iedge_mean = nullptr;	×
64	const data_flow::Memlet *iedge_var = nullptr;	×
65	const data_flow::Memlet *oedge_output = nullptr;	×
66	for (const auto &edge : dataflow.in_edges(*this)) {	×
67	if (edge.dst_conn() == "X") {	×
68	iedge_input = &edge;	×
69	} else if (edge.dst_conn() == "Scale") {	×
70	iedge_scale = &edge;	×
71	} else if (edge.dst_conn() == "B") {	×
72	iedge_bias = &edge;	×
73	} else if (edge.dst_conn() == "input_mean") {	×
74	iedge_mean = &edge;	×
75	} else if (edge.dst_conn() == "input_var") {	×
76	iedge_var = &edge;	×
77	}	×
78	}
79	for (const auto &edge : dataflow.out_edges(*this)) {	×
80	if (edge.src_conn() == "Y") {	×
81	oedge_output = &edge;	×
82	}	×
83	}
84	if (!iedge_input \|\| !iedge_scale \|\| !iedge_bias \|\| !iedge_mean \|\| !iedge_var \|\| !oedge_output) return false;	×
85
86	std::string input_name = static_cast<const data_flow::AccessNode &>(iedge_input->src()).data();	×
87	std::string scale_name = static_cast<const data_flow::AccessNode &>(iedge_scale->src()).data();	×
88	std::string bias_name = static_cast<const data_flow::AccessNode &>(iedge_bias->src()).data();	×
89	std::string mean_name = static_cast<const data_flow::AccessNode &>(iedge_mean->src()).data();	×
90	std::string var_name = static_cast<const data_flow::AccessNode &>(iedge_var->src()).data();	×
91	std::string output_name = static_cast<const data_flow::AccessNode &>(oedge_output->dst()).data();	×
92
93	// Create new sequence before
94	auto &new_sequence = builder.add_sequence_before(parent, block, transition.assignments(), block.debug_info());	×
95	structured_control_flow::Sequence *last_scope = &new_sequence;	×
96
UNCOV 97	std::vector<symbolic::Expression> loop_syms;	×
98	structured_control_flow::Map *last_map = nullptr;	×
NEW 99	for (size_t d = 0; d < this->shape_.size(); ++d) {	×
100	std::string indvar_str = builder.find_new_name("_i");	×
101	builder.add_container(indvar_str, types::Scalar(types::PrimitiveType::UInt64));	×
102	auto indvar = symbolic::symbol(indvar_str);	×
NEW 103	auto init = symbolic::zero();	×
104	auto update = symbolic::add(indvar, symbolic::one());	×
NEW 105	auto cond = symbolic::Lt(indvar, this->shape_[d]);	×
106	last_map = &builder.add_map(	×
107	*last_scope,	×
UNCOV 108	indvar,	×
UNCOV 109	cond,	×
UNCOV 110	init,	×
UNCOV 111	update,	×
112	structured_control_flow::ScheduleType_Sequential::create(),	×
113	{},	×
114	block.debug_info()	×
115	);
116	last_scope = &last_map->root();	×
117	loop_syms.push_back(indvar);	×
118	}	×
119
120	// Create normalization block
121	auto &norm_block = builder.add_block(*last_scope);	×
122
123	// Create access nodes for normalization
124	auto &input_access_norm = builder.add_access(norm_block, input_name);	×
125	auto &scale_access_norm = builder.add_access(norm_block, scale_name);	×
126	auto &bias_access_norm = builder.add_access(norm_block, bias_name);	×
127	auto &mean_access_norm = builder.add_access(norm_block, mean_name);	×
128	auto &var_access_norm = builder.add_access(norm_block, var_name);	×
129	auto &output_access_norm = builder.add_access(norm_block, output_name);	×
130
131	// Add epsilon to variance and compute standard deviation
132	auto &add_epsilon_tasklet =	×
133	builder.add_tasklet(norm_block, data_flow::TaskletCode::add, "_out", {"_in1", epsilon_});	×
134	auto &var_eps_access = builder.add_access(norm_block, builder.find_new_name("_var_eps"));	×
135	builder.add_computational_memlet(	×
136	norm_block, var_access_norm, add_epsilon_tasklet, "_in1", loop_syms, iedge_var->base_type()	×
137	);
138	builder	×
139	.add_computational_memlet(norm_block, add_epsilon_tasklet, "_out", var_eps_access, {}, iedge_var->base_type());	×
140
141	auto &sqrt_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::sqrt, "_out", {"_in"});	×
142	auto &std_dev_access = builder.add_access(norm_block, builder.find_new_name("_std_dev"));	×
143	builder.add_computational_memlet(norm_block, var_eps_access, sqrt_tasklet, "_in", {}, iedge_var->base_type());	×
144	builder.add_computational_memlet(norm_block, sqrt_tasklet, "_out", std_dev_access, {}, iedge_var->base_type());	×
145
146	// Normalize: (x - mean) / std_dev
147	auto &sub_norm_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::sub, "_out", {"_in1", "_in2"});	×
148	auto &centered_access = builder.add_access(norm_block, builder.find_new_name("_centered"));	×
149	builder.add_computational_memlet(	×
150	norm_block, input_access_norm, sub_norm_tasklet, "_in1", loop_syms, iedge_input->base_type()	×
151	);
152	builder.add_computational_memlet(	×
153	norm_block, mean_access_norm, sub_norm_tasklet, "_in2", loop_syms, iedge_mean->base_type()	×
154	);
155	builder	×
156	.add_computational_memlet(norm_block, sub_norm_tasklet, "_out", centered_access, {}, iedge_input->base_type());	×
157
158	auto &div_norm_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::div, "_out", {"_in1", "_in2"});	×
159	auto &normalized_access = builder.add_access(norm_block, builder.find_new_name("_normalized"));	×
160	builder	×
161	.add_computational_memlet(norm_block, centered_access, div_norm_tasklet, "_in1", {}, iedge_input->base_type());	×
162	builder	×
163	.add_computational_memlet(norm_block, std_dev_access, div_norm_tasklet, "_in2", loop_syms, iedge_var->base_type());	×
164	builder	×
165	.add_computational_memlet(norm_block, div_norm_tasklet, "_out", normalized_access, {}, iedge_input->base_type());	×
166
167	// Apply scale and bias: scale * normalized + bias
168	auto &mul_scale_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::mul, "_out", {"_in1", "_in2"});	×
169	auto &scaled_access = builder.add_access(norm_block, builder.find_new_name("_scaled"));	×
170	builder	×
171	.add_computational_memlet(norm_block, normalized_access, mul_scale_tasklet, "_in1", {}, iedge_input->base_type());	×
172	builder.add_computational_memlet(	×
173	norm_block, scale_access_norm, mul_scale_tasklet, "_in2", loop_syms, iedge_scale->base_type()	×
174	);
175	builder.add_computational_memlet(norm_block, mul_scale_tasklet, "_out", scaled_access, {}, iedge_input->base_type());	×
176
177	auto &add_bias_tasklet = builder.add_tasklet(norm_block, data_flow::TaskletCode::add, "_out", {"_in1", "_in2"});	×
178	builder.add_computational_memlet(norm_block, scaled_access, add_bias_tasklet, "_in1", {}, iedge_input->base_type());	×
179	builder.add_computational_memlet(	×
180	norm_block, bias_access_norm, add_bias_tasklet, "_in2", loop_syms, iedge_bias->base_type()	×
181	);
182	builder.add_computational_memlet(	×
183	norm_block, add_bias_tasklet, "_out", output_access_norm, loop_syms, oedge_output->base_type()	×
184	);
185
186	// Cleanup old block
187	builder.remove_memlet(block, *iedge_input);	×
188	builder.remove_memlet(block, *iedge_scale);	×
189	if (iedge_bias) {	×
190	builder.remove_memlet(block, *iedge_bias);	×
191	}	×
192	builder.remove_memlet(block, *iedge_mean);	×
193	builder.remove_memlet(block, *iedge_var);	×
194	builder.remove_memlet(block, *oedge_output);	×
195	builder.remove_node(block, *this);	×
196	builder.remove_child(parent, index + 1);	×
197
198	return true;	×
199	}	×
200
201	std::unique_ptr<data_flow::DataFlowNode> BatchNormalizationNode::
202	clone(size_t element_id, const graph::Vertex vertex, data_flow::DataFlowGraph &parent) const {	×
203	return std::unique_ptr<data_flow::DataFlowNode>(	×
NEW 204	new BatchNormalizationNode(element_id, this->debug_info(), vertex, parent, this->shape_, axis_, epsilon_)	×
205	);
206	}	×
207
208	nlohmann::json BatchNormalizationNodeSerializer::serialize(const data_flow::LibraryNode &library_node) {	×
209	const BatchNormalizationNode &node = static_cast<const BatchNormalizationNode &>(library_node);	×
210	nlohmann::json j;	×
211
212	j["code"] = node.code().value();	×
213	j["axis"] = node.axis();	×
214	j["epsilon"] = node.epsilon();	×
215
NEW 216	serializer::JSONSerializer serializer;	×
NEW 217	j["shape"] = nlohmann::json::array();	×
NEW 218	for (auto &dim : node.shape()) {	×
NEW 219	j["shape"].push_back(serializer.expression(dim));	×
220	}
221
222	return j;	×
223	}	×
224
225	data_flow::LibraryNode &BatchNormalizationNodeSerializer::deserialize(	×
226	const nlohmann::json &j, builder::StructuredSDFGBuilder &builder, structured_control_flow::Block &parent
227	) {
228	auto code = j["code"].get<std::string>();	×
229	if (code != LibraryNodeType_BatchNormalization.value()) {	×
230	throw std::runtime_error("Invalid library node code");	×
231	}
232
233	sdfg::serializer::JSONSerializer serializer;	×
234	DebugInfo debug_info = serializer.json_to_debug_info(j["debug_info"]);	×
235
NEW 236	std::vector<symbolic::Expression> shape;	×
NEW 237	for (const auto &dim : j["shape"]) {	×
NEW 238	shape.push_back(symbolic::parse(dim.get<std::string>()));	×
239	}
240
241	auto axis = j["axis"].get<int>();	×
242	auto epsilon = j["epsilon"].get<std::string>();	×
243
NEW 244	return builder.add_library_node<BatchNormalizationNode>(parent, debug_info, shape, axis, epsilon);	×
245	}	×
246
247	} // namespace ml
248	} // namespace math
249	} // namespace sdfg

daisytuner / sdfglib / 17697974118

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous