Skip to content

Commit

Permalink
make log less verbose
Browse files Browse the repository at this point in the history
  • Loading branch information
@daquexian
daquexian committed Jul 11, 2019
1 parent b5739d5 commit efa45e7
Show file tree
Hide file tree
Showing 2 changed files with 34 additions and 35 deletions.
1 change: 0 additions & 1 deletion dnnlibrary/src/DaqReader.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -183,7 +183,6 @@ void AddOutputs(const DNN::Model &model, ModelBuilder &builder) {
return;
}
for (const auto &output : *model.outputs()) {
PNT(output->str());
css output_name = output->str();
builder.AddOutput(output_name);
}
Expand Down
68 changes: 34 additions & 34 deletions tools/onnx2daq/OnnxConverter.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -228,7 +228,7 @@ void OnnxConverter::AddConv(const string &input_name,
throw std::invalid_argument(
"Both dilations and strides > 1 is not supported for now");
}
LOG(INFO) << "Dilations of conv: " << dilations << ", converting..";
VLOG(5) << "Dilations of conv: " << dilations << ", converting..";
const auto s2b_name = input_name + "_s2b";
const auto im_name = input_name + "_conv_imm";
const auto b2s_name = input_name + "_b2s";
Expand All @@ -240,7 +240,7 @@ void OnnxConverter::AddConv(const string &input_name,
new_pads[3] = (input_shape[2] + pads[3] + (dilations[1] - 1)) /
dilations[1] * dilations[1] -
input_shape[2];
LOG(INFO) << input_shape << ", " << pads << ", " << dilations << ", "
VLOG(5) << input_shape << ", " << pads << ", " << dilations << ", "
<< new_pads;
// Why "AllowShortBlocksOnASingleLine: false" doesn't work on it?
// clang-format off
Expand Down Expand Up @@ -280,11 +280,11 @@ void OnnxConverter::AddConv(const string &input_name,

const auto &onnx_weight = onnx_tensors_.at(ori_weight_name);
if (group == 1) {
LOG(INFO) << "Vanilla conv";
VLOG(5) << "Vanilla conv";
AddLayerConvImpl(input_name, ori_weight_name, bias_name, pads, strides,
output_name);
} else if (onnx_weight.shape[1] == 1) { // depthwise
LOG(INFO) << "Depthwise conv";
VLOG(5) << "Depthwise conv";
AddLayerDepthwiseConvImpl(input_name, ori_weight_name, bias_name, pads,
strides, 1, output_name);
} else {
Expand Down Expand Up @@ -1171,10 +1171,10 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
const auto &node = model_proto_.graph().node(i);
NodeAttrHelper helper(node);
const auto &op = node.op_type();
LOG(INFO) << "Node " << node.name();
VLOG(5) << "Node " << node.name();
if (std::find(skipped_act_.begin(), skipped_act_.end(), i) !=
skipped_act_.end()) {
LOG(INFO) << "Skip layer " << node.name();
VLOG(5) << "Skip layer " << node.name();
continue;
}
if (has_reshape && op != "Gemm") {
Expand All @@ -1183,7 +1183,7 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
"for now");
}
if (op == "Conv") {
LOG(INFO) << "Start converting Conv";
VLOG(5) << "Start converting Conv";
const auto strides = helper.get("strides", vector<int>{1, 1});
const auto pads = helper.get("pads", vector<int>{0, 0, 0, 0});
const auto dilations = helper.get("dilations", vector<int>{1, 1});
Expand All @@ -1199,10 +1199,10 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
const auto ori_weight_name = m(node.input(1));
AddConv(m(node.input(0)), strides, pads, dilations, group,
ori_weight_name, bias_name, m(node.output(0)));
LOG(INFO) << "Converting Conv completed";
VLOG(5) << "Converting Conv completed";
} else if (op == "AveragePool" || op == "MaxPool" ||
op == "GlobalAveragePool" || op == "GlobalMaxPool") {
LOG(INFO) << "Start converting Pool";
VLOG(5) << "Start converting Pool";
const auto input_name = m(node.input(0));
const auto output_name = m(node.output(0));
vector<int> strides, pads, kernel_shape;
Expand Down Expand Up @@ -1234,16 +1234,16 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
CHECK_EQ(strides.size(), 2ul);
AddLayerPool(op, input_name, kernel_shape, pads, strides,
output_name);
LOG(INFO) << "Converting Pool completed";
VLOG(5) << "Converting Pool completed";
} else if (op == "Relu") {
LOG(INFO) << "Start converting Relu";
VLOG(5) << "Start converting Relu";
const auto input_name = m(node.input(0));
const auto output_name = m(node.output(0));
AddLayerReLU(input_name, output_name);
LOG(INFO) << "Converting Relu completed";
VLOG(5) << "Converting Relu completed";

} else if (op == "PRelu") {
LOG(INFO) << "Start converting PRelu";
VLOG(5) << "Start converting PRelu";
const auto input_name = m(node.input(0));
const auto slope_name = m(node.input(1));
const auto output_name = m(node.output(0));
Expand All @@ -1262,23 +1262,23 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
AddLayerMul(imm3_name, -1.f, imm4_name);
AddLayerAdd(imm1_name, imm4_name, output_name);
// TODO:
LOG(INFO) << "Converting PRelu completed";
VLOG(5) << "Converting PRelu completed";
} else if (op == "Add") {
LOG(INFO) << "Start converting Add";
VLOG(5) << "Start converting Add";
const auto input1_name = m(node.input(0));
const auto input2_name = m(node.input(1));
const auto output_name = m(node.output(0));
AddLayerAdd(input1_name, input2_name, output_name);
LOG(INFO) << "Converting Add completed";
VLOG(5) << "Converting Add completed";
} else if (op == "Mul") {
LOG(INFO) << "Start converting Mul";
VLOG(5) << "Start converting Mul";
const auto input1_name = m(node.input(0));
const auto input2_name = m(node.input(1));
const auto output_name = m(node.output(0));
AddLayerMul(input1_name, input2_name, output_name);
LOG(INFO) << "Converting Mul completed";
VLOG(5) << "Converting Mul completed";
} else if (op == "Gemm") {
LOG(INFO) << "Start converting Gemm";
VLOG(5) << "Start converting Gemm";
const auto input_name = m(node.input(0));
const auto weight_name = m(node.input(1));
const auto output_name = m(node.output(0));
Expand All @@ -1299,15 +1299,15 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
"supported.");
}
has_reshape = false;
LOG(INFO) << "Converting Gemm completed";
VLOG(5) << "Converting Gemm completed";
} else if (op == "Softmax") {
LOG(INFO) << "Start converting Softmax";
VLOG(5) << "Start converting Softmax";
const auto input_name = m(node.input(0));
const auto output_name = m(node.output(0));
AddLayerSoftmax(input_name, output_name);
LOG(INFO) << "Converting Softmax completed";
VLOG(5) << "Converting Softmax completed";
} else if (op == "Concat") {
LOG(INFO) << "Start converting Concat";
VLOG(5) << "Start converting Concat";
vector<std::string> concat_inputs_str;
for (const auto &onnx_input : node.input()) {
concat_inputs_str.push_back(m(onnx_input));
Expand All @@ -1317,14 +1317,14 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
const auto output_name = m(node.output(0));
AddLayerConcat(concat_inputs_str, axis_nchw_to_nhwc[axis],
output_name);
LOG(INFO) << "Converting Concat completed";
VLOG(5) << "Converting Concat completed";
} else if (op == "Dropout") {
LOG(INFO) << "Start converting Dropout";
VLOG(5) << "Start converting Dropout";
SetIdentity(node.input(0), node.output(0));
LOG(INFO) << "Converting Dropout completed";
VLOG(5) << "Converting Dropout completed";

} else if (op == "BatchNormalization") {
LOG(INFO) << "Start converting BatchNormalization";
VLOG(5) << "Start converting BatchNormalization";
DNN_ASSERT(node.output_size() == 1,
"Your onnx model may be in training mode, please export "
"it in test mode.")
Expand Down Expand Up @@ -1370,14 +1370,14 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
AddLayerMul(input_name, tensor_a_name, tensor_imm_product_name);
AddLayerAdd(tensor_imm_product_name, tensor_b_name, output_name);

LOG(INFO) << "Converting BatchNormalization completed";
VLOG(5) << "Converting BatchNormalization completed";
} else if (op == "Reshape") {
LOG(INFO) << "Start converting Reshape";
VLOG(5) << "Start converting Reshape";
has_reshape = true;
SetIdentity(node.input(0), node.output(0));
LOG(INFO) << "Converting Reshape completed";
VLOG(5) << "Converting Reshape completed";
} else if (op == "LRN") {
LOG(INFO) << "Start converting LRN";
VLOG(5) << "Start converting LRN";
if (!helper.has_attr("size")) {
throw std::invalid_argument(
"Invalid ONNX model, attribute \"size\" is required in "
Expand All @@ -1395,7 +1395,7 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,
// that of NNAPI LRN
AddLayerLRN(node.input(0), radius, bias, alpha, beta,
node.output(0));
LOG(INFO) << "Converting LRN completed";
VLOG(5) << "Converting LRN completed";
} else {
throw std::invalid_argument("Unsupported operator " + op);
}
Expand Down Expand Up @@ -1423,8 +1423,8 @@ void OnnxConverter::Convert(const ONNX_NAMESPACE::ModelProto &model_proto,

builder_.Finish(flat_model);

LOG(INFO) << "Shapes: ";
LOG(INFO) << shaper_;
VLOG(5) << "Shapes: ";
VLOG(5) << shaper_;

Clear();
}
Expand Down

0 comments on commit efa45e7

Please sign in to comment.