Skip to content

GitLab

Commit 1d5d035c authored by Shucai Xiao's avatar Shucai Xiao
Browse files

clang format

parent ea9776f5
Hide whitespace changes
Inline Side-by-side
Showing with 60 additions and 40 deletions
src/quantization.cpp
View file @ 1d5d035c
...@@ -153,15 +153,16 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names) ...@@ -153,15 +153,16 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names)
// operation, if it has 3 inputs, then the last one should // operation, if it has 3 inputs, then the last one should
// be converted to int32_type // be converted to int32_type
shape::type_t quant_type = shape::int8_type; shape::type_t quant_type = shape::int8_type;
auto param = int8_param[param_index++]; auto param = int8_param[param_index++];
if(ins->name() == "dot" and inputs.size() == 3 and input == inputs.back()) if(ins->name() == "dot" and inputs.size() == 3 and input == inputs.back())
{ {
quant_type = shape::int32_type; quant_type = shape::int32_type;
} }
auto s = input->get_shape(); auto s = input->get_shape();
if((s.type() == shape::float_type || s.type() == shape::double_type || if((s.type() == shape::float_type || s.type() == shape::double_type ||
s.type() == shape::int32_type) && s.type() != quant_type) s.type() == shape::int32_type) &&
s.type() != quant_type)
{ {
// if the input is a convert operator, uses its input // if the input is a convert operator, uses its input
// as its current input // as its current input
...@@ -204,66 +205,74 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names) ...@@ -204,66 +205,74 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names)
// equal)", we need additional calculation for the adjustment // equal)", we need additional calculation for the adjustment
if(ins->name() == "dot") if(ins->name() == "dot")
{ {
auto dot_op = any_cast<op::dot>(ins->get_operator()); auto dot_op = any_cast<op::dot>(ins->get_operator());
float new_alpha = dot_op.alpha / (int8_param[0].first * int8_param[1].first); float new_alpha = dot_op.alpha / (int8_param[0].first * int8_param[1].first);
float new_beta = dot_op.beta; float new_beta = dot_op.beta;
// We need additional checking about the quant_alpha value. If // We need additional checking about the quant_alpha value. If
// abs(quant_alpha) > 50 (some tmp value set here), we can convert // abs(quant_alpha) > 50 (some tmp value set here), we can convert
// it to an integer as the new_alpha in the quant_dot // it to an integer as the new_alpha in the quant_dot
float threshold = 50.0f; float threshold = 50.0f;
if (fabs(new_alpha) >= threshold && fabs(new_beta) >= threshold) if(fabs(new_alpha) >= threshold && fabs(new_beta) >= threshold)
{ {
int32_t quant_alpha = static_cast<int32_t>(new_alpha); int32_t quant_alpha = static_cast<int32_t>(new_alpha);
int32_t quant_beta = static_cast<int32_t>(new_beta); int32_t quant_beta = static_cast<int32_t>(new_beta);
shape quant_shape = compute_shape(op::quant_dot{1, 0}, converted_inputs); shape quant_shape = compute_shape(op::quant_dot{1, 0}, converted_inputs);
if (quant_shape.type() == orig_type) if(quant_shape.type() == orig_type)
{ {
prog.replace_instruction(ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs); prog.replace_instruction(
ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs);
} }
else else
{ {
auto quant_dot = prog.insert_instruction(ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs); auto quant_dot = prog.insert_instruction(
ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs);
prog.replace_instruction(ins, op::convert{orig_type}, quant_dot); prog.replace_instruction(ins, op::convert{orig_type}, quant_dot);
} }
} }
// only alpha can be quantized, quantization of beta will cause // only alpha can be quantized, quantization of beta will cause
// big error, so we have to manually do the multiplication and // big error, so we have to manually do the multiplication and
// addition // addition
else if (fabs(new_alpha) >= threshold) else if(fabs(new_alpha) >= threshold)
{ {
int32_t quant_alpha = static_cast<int32_t>(new_alpha); int32_t quant_alpha = static_cast<int32_t>(new_alpha);
int32_t quant_beta = 0; int32_t quant_beta = 0;
if (orig_type == shape::int32_type) if(orig_type == shape::int32_type)
{ {
if (inputs.size() == 2 or dot_op.beta == 0.0f) if(inputs.size() == 2 or dot_op.beta == 0.0f)
{ {
prog.replace_instruction(ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs); prog.replace_instruction(
ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs);
} }
// if there are 3 inputs, we need to consider the third argument // if there are 3 inputs, we need to consider the third argument
else else
{ {
auto q_dot = prog.insert_instruction(ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs); auto q_dot = prog.insert_instruction(
ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs);
std::vector<float> vec_beta(q_dot->get_shape().elements(), dot_op.beta); std::vector<float> vec_beta(q_dot->get_shape().elements(), dot_op.beta);
auto l_beta = prog.add_literal(literal{orig_type, vec_beta}); auto l_beta = prog.add_literal(literal{orig_type, vec_beta});
auto beta_c = prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back()); auto beta_c =
prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back());
prog.replace_instruction(ins, op::add{}, q_dot, beta_c); prog.replace_instruction(ins, op::add{}, q_dot, beta_c);
} }
} }
else else
{ {
if (inputs.size() == 2 or dot_op.beta == 0.0f) if(inputs.size() == 2 or dot_op.beta == 0.0f)
{ {
auto q_dot = prog.insert_instruction(ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs); auto q_dot = prog.insert_instruction(
ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs);
prog.replace_instruction(ins, op::convert{orig_type}, q_dot); prog.replace_instruction(ins, op::convert{orig_type}, q_dot);
} }
// if there are 3 inputs, we need to consider the third argument // if there are 3 inputs, we need to consider the third argument
else else
{ {
auto q_dot = prog.insert_instruction(ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs); auto q_dot = prog.insert_instruction(
ins, op::quant_dot{quant_alpha, quant_beta}, converted_inputs);
auto oq_dot = prog.insert_instruction(ins, op::convert{orig_type}, q_dot); auto oq_dot = prog.insert_instruction(ins, op::convert{orig_type}, q_dot);
std::vector<float> vec_beta(q_dot->get_shape().elements(), dot_op.beta); std::vector<float> vec_beta(q_dot->get_shape().elements(), dot_op.beta);
auto l_beta = prog.add_literal(literal{oq_dot->get_shape(), vec_beta}); auto l_beta = prog.add_literal(literal{oq_dot->get_shape(), vec_beta});
auto beta_c = prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back()); auto beta_c =
prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back());
prog.replace_instruction(ins, op::add{}, q_dot, beta_c); prog.replace_instruction(ins, op::add{}, q_dot, beta_c);
} }
} }
...@@ -272,10 +281,10 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names) ...@@ -272,10 +281,10 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names)
{ {
auto q_dot = prog.insert_instruction(ins, op::quant_dot{1, 0}, converted_inputs); auto q_dot = prog.insert_instruction(ins, op::quant_dot{1, 0}, converted_inputs);
std::vector<float> vec_alpha(q_dot->get_shape().elements(), new_alpha); std::vector<float> vec_alpha(q_dot->get_shape().elements(), new_alpha);
if (orig_type == shape::int32_type) if(orig_type == shape::int32_type)
{ {
auto l_alpha = prog.add_literal(literal(ins->get_shape(), vec_alpha)); auto l_alpha = prog.add_literal(literal(ins->get_shape(), vec_alpha));
if (converted_inputs.size() == 2 or dot_op.beta == 0.0f) if(converted_inputs.size() == 2 or dot_op.beta == 0.0f)
{ {
prog.replace_instruction(ins, op::mul{}, l_alpha, q_dot); prog.replace_instruction(ins, op::mul{}, l_alpha, q_dot);
} }
...@@ -283,17 +292,18 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names) ...@@ -283,17 +292,18 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names)
else else
{ {
std::vector<float> vec_beta(ins->get_shape().elements(), new_beta); std::vector<float> vec_beta(ins->get_shape().elements(), new_beta);
auto l_beta = prog.add_literal(literal(ins->get_shape(), vec_beta)); auto l_beta = prog.add_literal(literal(ins->get_shape(), vec_beta));
auto alpha_ab = prog.insert_instruction(ins, op::mul{}, l_alpha, q_dot); auto alpha_ab = prog.insert_instruction(ins, op::mul{}, l_alpha, q_dot);
auto beta_c = prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back()); auto beta_c =
prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back());
prog.replace_instruction(ins, op::add{}, alpha_ab, beta_c); prog.replace_instruction(ins, op::add{}, alpha_ab, beta_c);
} }
} }
else else
{ {
auto oq_dot = prog.insert_instruction(ins, op::convert{orig_type}, q_dot); auto oq_dot = prog.insert_instruction(ins, op::convert{orig_type}, q_dot);
auto l_alpha = prog.add_literal(literal(ins->get_shape(), vec_alpha)); auto l_alpha = prog.add_literal(literal(ins->get_shape(), vec_alpha));
if (converted_inputs.size() == 2 or dot_op.beta == 0.0f) if(converted_inputs.size() == 2 or dot_op.beta == 0.0f)
{ {
prog.replace_instruction(ins, op::mul{}, l_alpha, oq_dot); prog.replace_instruction(ins, op::mul{}, l_alpha, oq_dot);
} }
...@@ -301,9 +311,10 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names) ...@@ -301,9 +311,10 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names)
else else
{ {
std::vector<float> vec_beta(ins->get_shape().elements(), new_beta); std::vector<float> vec_beta(ins->get_shape().elements(), new_beta);
auto l_beta = prog.add_literal(literal(ins->get_shape(), vec_beta)); auto l_beta = prog.add_literal(literal(ins->get_shape(), vec_beta));
auto alpha_ab = prog.insert_instruction(ins, op::mul{}, l_alpha, oq_dot); auto alpha_ab = prog.insert_instruction(ins, op::mul{}, l_alpha, oq_dot);
auto beta_c = prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back()); auto beta_c =
prog.insert_instruction(ins, op::mul{}, l_beta, inputs.back());
prog.replace_instruction(ins, op::add{}, alpha_ab, beta_c); prog.replace_instruction(ins, op::add{}, alpha_ab, beta_c);
} }
} }
...@@ -323,23 +334,32 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names) ...@@ -323,23 +334,32 @@ void quantize_int8(program& prog, const std::vector<std::string>& ins_names)
shape quant_shape = compute_shape(op::quant_convolution{}, converted_inputs); shape quant_shape = compute_shape(op::quant_convolution{}, converted_inputs);
std::vector<float> vec_factor(quant_shape.elements(), adjust_factor); std::vector<float> vec_factor(quant_shape.elements(), adjust_factor);
auto fl = prog.add_literal(literal{{orig_type, quant_shape.lens()}, vec_factor}); auto fl = prog.add_literal(literal{{orig_type, quant_shape.lens()}, vec_factor});
if (quant_shape.type() == orig_type) if(quant_shape.type() == orig_type)
{ {
if (adjust_factor == 1.0f) if(adjust_factor == 1.0f)
{ {
prog.replace_instruction(ins, op::quant_convolution{padding, stride, dilation, padding_mode, group}, converted_inputs); prog.replace_instruction(
ins,
op::quant_convolution{padding, stride, dilation, padding_mode, group},
converted_inputs);
} }
else else
{ {
auto quant_conv = prog.replace_instruction(ins, op::quant_convolution{padding, stride, dilation, padding_mode, group}, converted_inputs); auto quant_conv = prog.replace_instruction(
ins,
op::quant_convolution{padding, stride, dilation, padding_mode, group},
converted_inputs);
prog.replace_instruction(ins, op::mul{}, quant_conv, fl); prog.replace_instruction(ins, op::mul{}, quant_conv, fl);
} }
} }
else else
{ {
auto quant_conv = prog.insert_instruction(ins, op::quant_convolution{padding, stride, dilation, padding_mode, group}, converted_inputs); auto quant_conv = prog.insert_instruction(
if (adjust_factor == 1.0f) ins,
op::quant_convolution{padding, stride, dilation, padding_mode, group},
converted_inputs);
if(adjust_factor == 1.0f)
{ {
prog.replace_instruction(ins, op::convert{orig_type}, quant_conv); prog.replace_instruction(ins, op::convert{orig_type}, quant_conv);
} }
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment