Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Netron ouptut fix #3778

Merged
merged 5 commits into from
Jan 24, 2025
Merged

Netron ouptut fix #3778

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
7ca8c89
Fix the value outputs and bugged string output in netron
CharlieL7 Jan 23, 2025
b270417
formatting
CharlieL7 Jan 23, 2025
bbcfdb0
Merge branch 'develop' into netron_ouptut_fix
CharlieL7 Jan 23, 2025
98bce5a
cppcheck suppress and licensing
CharlieL7 Jan 24, 2025
2d62c3b
Merge branch 'netron_ouptut_fix' of github.com:ROCmSoftwarePlatform/A…
CharlieL7 Jan 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
54 changes: 33 additions & 21 deletions src/netron_output.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
/*
* The MIT License (MIT)
*
* Copyright (c) 2015-2024 Advanced Micro Devices, Inc. All rights reserved.
* Copyright (c) 2015-2025 Advanced Micro Devices, Inc. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
Expand Down Expand Up @@ -63,12 +63,19 @@ int get_onnx_type(shape::type_t s_type)

auto make_attribute(const migraphx::value& val)
{
value attribute;
value attribute = value(std::unordered_map<std::string, value>());
attribute["name"] = val.get_key();
auto val_string = val.to<std::string>();
val_string = val_string.substr(val_string.find(":") + 1);
attribute["s"] = base64_encode(val_string);
attribute["type"] = "STRING";
std::string sub_str = val.get_key() + ":";
auto find_key = val_string.find(sub_str);
if(find_key != std::string::npos)
{
val_string = val_string.substr(find_key + sub_str.length() + 1);
}
// TODO: doesn't work for some reason with Netron now
// attribute["s"] = base64_encode(val_string);
// attribute["type"] = "STRING";
attribute["docString"] = val_string;
return attribute;
}

Expand All @@ -78,7 +85,7 @@ auto make_onnx_json_node(instruction_ref ins,
{
value node;
// TODO add support for module inputs
value input_arr;
value input_arr = value({});
for(instruction_ref input_ins : ins->inputs())
{
auto name = input_ins->name();
Expand All @@ -96,7 +103,7 @@ auto make_onnx_json_node(instruction_ref ins,
input_arr.push_back(ins_uids.at(input_ins) + "->" + ins_uids.at(ins));
}
}
value output_arr;
value output_arr = value({});
for(instruction_ref output_ins : ins->outputs())
{
if(output_ins->name() == "@return")
Expand All @@ -108,15 +115,15 @@ auto make_onnx_json_node(instruction_ref ins,
output_arr.push_back(ins_uids.at(ins) + "->" + ins_uids.at(output_ins));
}
}
node["input"] = input_arr;
node["input"] = input_arr;
node["output"] = output_arr;
node["name"] = ins_uids.at(ins);
node["opType"] = ins->name();
value op_attribute_arr;
value op_attribute_arr = value({});
auto op_value = ins->get_operator().to_value();
std::for_each(op_value.begin(), op_value.end(), [&](auto v) {
const std::string& attr_key = v.get_key();
if(v.is_binary())
if(v.is_binary() or attr_key == "code_object")
{
return;
}
Expand Down Expand Up @@ -151,17 +158,17 @@ auto make_onnx_json_literal(instruction_ref ins,
auto make_onnx_json_shape(const shape& s)
{
value ret;
value dim;
auto shape_lens = s.lens();
std::transform(shape_lens.begin(),
shape_lens.end(),
std::back_inserter(dim),
[](std::size_t len) { return len; });
value dim = value({});
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually value::object{} would be cleaner here.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, didn't realize that's how it worked.

for(std::size_t len : s.lens())
{
// cppcheck-suppress useStlAlgorithm
dim.push_back({{"dimValue", len}});
}
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does the std::back_inserter not work with value class?

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I see, its supposed to be an object.

ret["dim"] = dim;
return ret;
}

// ONNX graph edges called "valuetype"
// ONNX graph edges called "valueType"
auto make_onnx_json_edge(instruction_ref ins,
instruction_ref out_ins,
std::unordered_map<instruction_ref, std::string> ins_uids)
Expand Down Expand Up @@ -207,8 +214,11 @@ std::unordered_map<instruction_ref, std::string> make_ins_uids(const module& mod

value make_graph(const module* mod)
{
value graph = {
{"node", {}}, {"initializer", {}}, {"input", {}}, {"output", {}}, {"valueInfo", {}}};
value graph = {{"node", value({})},
{"initializer", value({})},
{"input", value({})},
{"output", value({})},
{"valueInfo", value({})}};
auto ins_uids = make_ins_uids(*mod);
for(auto ins = mod->begin(); ins != mod->end(); ++ins)
{
Expand Down Expand Up @@ -251,15 +261,17 @@ std::string make_netron_output(const program& prog)
{
value output;
auto prog_value = prog.to_value();
output["irVersion"] = prog_value.at("version").to<std::string>();
// ONNX IR version 6
// TODO: investigate sure how this affects things
output["irVersion"] = 6;
output["producerName"] = "AMDMIGraphX";
output["producerVersion"] = prog_value.at("migraphx_version").to<std::string>();
for(auto& mod : prog.get_modules())
{
auto graph = make_graph(mod);
output["graph"] = graph;
}
return to_json_string(output);
return to_pretty_json_string(output, 4);
}

} // namespace MIGRAPHX_INLINE_NS
Expand Down
Loading