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

[Relay][Frontend][Onnx] Auto extract onnx input shapes when possible. #7115

Merged
merged 2 commits into from
Dec 16, 2020
Merged

[Relay][Frontend][Onnx] Auto extract onnx input shapes when possible. #7115

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
389820c
Auto extract onnx input shapes when possible.
jwfromm Dec 15, 2020
895954f
Remove shape dict definition in tvmc.
jwfromm Dec 15, 2020
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
11 changes: 1 addition & 10 deletions python/tvm/driver/tvmc/frontends.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -161,16 +161,7 @@ def load(self, path):
# pylint: disable=E1101
model = onnx.load(path)

# pylint: disable=E1101
name = model.graph.input[0].name

# pylint: disable=E1101
proto_shape = model.graph.input[0].type.tensor_type.shape.dim
shape = [d.dim_value for d in proto_shape]

shape_dict = {name: shape}

return relay.frontend.from_onnx(model, shape_dict)
return relay.frontend.from_onnx(model)


class TensorflowFrontend(Frontend):
Expand Down
37 changes: 19 additions & 18 deletions python/tvm/relay/frontend/onnx.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -111,15 +111,20 @@ def get_type(elem_type):
def get_info(info_proto):
"""Extract the shape from a ValueInfoProto."""
shape = []
shape_name = []
for dim in info_proto.type.tensor_type.shape.dim:
name = dim.dim_param
value = dim.dim_value
if value is None:
value = _ty.Any
if value is None or value == 0:
value = _ty.Any()
shape_name.append(name)
else:
shape_name.append(value)
shape.append(value)

name = info_proto.name
dtype = get_type(info_proto.type.tensor_type.elem_type)
return name, shape, dtype
return name, shape, dtype, shape_name


def dimension_picker(prefix, suffix=""):
Expand Down Expand Up @@ -2185,7 +2190,7 @@ def get_var(name, val, scan=False):
scan_output_vars = []
scan_output_init = []
for i in range(num_scan_outputs):
name, shape, dtype = get_info(body.output[i + 1 + num_deps])
name, shape, dtype, _ = get_info(body.output[i + 1 + num_deps])
scan_output_vars.append(_expr.var(name, shape=([_ty.Any()] + shape), dtype=dtype))
scan_output_init.append(_op.reshape(_expr.const([]), [0] + shape))

Expand Down Expand Up @@ -2829,8 +2834,7 @@ def from_onnx(self, graph, opset, freeze_params=False, get_output_expr=False):
for i in graph.input:
# from onnx v0.2, GraphProto.input has type ValueInfoProto,
# and the name is 'i.name'
i_name = self._parse_value_proto(i)
d_type = self._parse_dtype(i, "float32")
i_name, i_shape, d_type, i_shape_name = get_info(i)
if i_name in self._params:
# i is a param instead of input
self._num_param += 1
Expand All @@ -2841,14 +2845,20 @@ def from_onnx(self, graph, opset, freeze_params=False, get_output_expr=False):
else:
self._num_input += 1
if i_name in self._shape:
tshape = self._shape[i_name]
i_shape = self._shape[i_name]
else:
raise ValueError("Must provide an input shape for `{0}`.".format(i_name))
if "?" in str(i_shape):
warning_msg = (
"Input %s has unknown dimension shapes: %s. "
"Specifying static values may improve performance"
% (i_name, str(i_shape_name))
)
warnings.warn(warning_msg)
if isinstance(self._dtype, dict):
dtype = self._dtype[i_name] if i_name in self._dtype else d_type
else:
dtype = d_type
self._nodes[i_name] = new_var(i_name, shape=tshape, dtype=dtype)
self._nodes[i_name] = new_var(i_name, shape=i_shape, dtype=dtype)
self._inputs[i_name] = self._nodes[i_name]
# get list of unsupported ops
convert_map = _get_convert_map(opset)
Expand Down Expand Up @@ -2935,15 +2945,6 @@ def _parse_value_proto(self, value_proto):
name = value_proto
return name

def _parse_dtype(self, value_proto, dtype):
"""Parse dtype."""
try:
from onnx.mapping import TENSOR_TYPE_TO_NP_TYPE

return TENSOR_TYPE_TO_NP_TYPE[value_proto.type.tensor_type.elem_type].name
except AttributeError:
return dtype

def _parse_array(self, tensor_proto):
np_array = get_numpy(tensor_proto).reshape(tuple(tensor_proto.dims))
return _nd.array(np_array)
Expand Down