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

Jump to bottom

Enable python debug runtime for exported network libraries #8793

Merged
merged 7 commits into from
Sep 2, 2021
Merged

Enable python debug runtime for exported network libraries #8793

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
54bdbcd
Add get_json method to graph_eceutor factory
apeskov Jun 29, 2021
bbf3397
Update Debugger runtime documentation for exported libraries
elvin-n Aug 19, 2021
472684d
Fix cpplint
elvin-n Aug 19, 2021
660a8fb
Change module get_json to get_graph_json, add test
elvin-n Aug 25, 2021
fe97c33
Fix get_graph_json test
elvin-n Aug 25, 2021
7b1f3f6
Change verificatino of llvm support in tet to decorator
elvin-n Aug 26, 2021
3f4f859
Fix sphinx warning in debugger.rst
elvin-n Aug 30, 2021
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
Prev Previous commit
Next Next commit
Change module get_json to get_graph_json, add test
  • Loading branch information
@elvin-n
elvin-n committed Aug 29, 2021
commit 660a8fbc960a811e7d7d6d8309b128762c15299a
7 changes: 3 additions & 4 deletions docs/dev/debugger.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,12 +123,12 @@ Example of loading the parameters
How to use Debugger?
***************************************

1. In ``config.cmake`` set the ``USE_GRAPH_RUNTIME_DEBUG`` flag to ``ON``
1. In ``config.cmake`` set the ``USE_PROFILER`` flag to ``ON``

::

# Whether enable additional graph debug functions
set(USE_GRAPH_RUNTIME_DEBUG ON)
set(USE_PROFILER ON)

2. Do 'make' tvm, so that it will make the ``libtvm_runtime.so``

Expand All @@ -154,8 +154,7 @@ How to use Debugger?

::
lib = tvm.runtime.load_module("network.so")
m = graph_executor.GraphModuleDebug(lib["debug_create"]("default", dev),
[dev], lib["get_json"](), dump_root="/tmp/tvmdbg")
m = graph_executor.create(lib["get_graph_json"](), lib, dev, dump_root="/tmp/tvmdbg")
# set inputs
m.set_input('data', tvm.nd.array(data.astype(dtype)))
m.set_input(**params)
Expand Down
3 changes: 2 additions & 1 deletion src/runtime/graph_executor/graph_executor_factory.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,9 +53,10 @@ PackedFunc GraphExecutorFactory::GetFunction(
}
*rv = this->ExecutorCreate(devices);
});
} else if (name == "get_json") {
} else if (name == "get_graph_json") {
return PackedFunc(
[sptr_to_self, this](TVMArgs args, TVMRetValue* rv) { *rv = this->graph_json_; });

} else if (name == "debug_create") {
return PackedFunc([sptr_to_self, this](TVMArgs args, TVMRetValue* rv) {
ICHECK_GE(args.size(), 2);
Expand Down
20 changes: 20 additions & 0 deletions tests/python/unittest/test_runtime_module_based_interface.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -90,6 +90,25 @@ def test_cpu():
tvm.testing.assert_allclose(out, verify(data), atol=1e-5)


def test_cpu_get_graph_json():
if not tvm.testing.device_enabled("llvm"):
print("Skip because llvm is not enabled")
return
Copy link
Contributor

Choose a reason for hiding this comment

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

Suggested change
def test_cpu_get_graph_json():
if not tvm.testing.device_enabled("llvm"):
print("Skip because llvm is not enabled")
return
@tvm.testing.requires_llvm
def test_cpu_get_graph_json():

Copy link
Contributor Author

Choose a reason for hiding this comment

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

fixed together with other tests in this script

mod, params = relay.testing.synthetic.get_workload()
with relay.build_config(opt_level=3):
complied_graph_lib = relay.build_module.build(mod, "llvm", params=params)
from tvm.contrib import utils

temp = utils.tempdir()
file_name = "deploy_lib.so"
path_lib = temp.relpath(file_name)
complied_graph_lib.export_library(path_lib)
loaded_lib = tvm.runtime.load_module(path_lib)
json = loaded_lib["get_graph_json"]()
assert isinstance(json, str) == True
assert json.find("tvmgen_default_fused_nn_softmax1") == 6312


@tvm.testing.requires_cuda
@tvm.testing.requires_gpu
def test_gpu():
Expand Down Expand Up @@ -619,3 +638,4 @@ def make_module(mod):
test_remove_package_params()
test_debug_graph_executor()
test_multiple_imported_modules()
test_cpu_get_graph_json()