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.

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

[Frontend][Tensorflow] SelectV2 and BroadcastArgs op support for tf2 models #7901

Merged
merged 5 commits into from
Apr 24, 2021
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
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
2 changes: 1 addition & 1 deletion 3rdparty/vta-hw
43 changes: 40 additions & 3 deletions python/tvm/relay/frontend/tensorflow.py
Original file line number Diff line number Diff line change
Expand Up @@ -1765,6 +1765,44 @@ def _impl(inputs, attr, params, mod):
return _impl


def _broadcast_args():
def _impl(inputs, attr, params, mod):
if isinstance(inputs[0], _expr.Var):
s0 = params[inputs[0].name_hint]
else:
s0 = _infer_value(inputs[0], params, mod)
if isinstance(inputs[1], _expr.Var):
s1 = params[inputs[1].name_hint]
else:
s1 = _infer_value(inputs[1], params, mod)
s0 = list(s0.asnumpy().reshape([-1]))
s1 = list(s1.asnumpy().reshape([-1]))
s0_size, s1_size = len(s0), len(s1)
from collections import deque
comaniac marked this conversation as resolved.
Show resolved Hide resolved
comaniac marked this conversation as resolved.
Show resolved Hide resolved

out = deque([])
for i in range(1, min(s0_size, s1_size) + 1):
if s0[s0_size - i] == s1[s1_size - i]:
out.appendleft(s0[s0_size - i])
elif s0[s0_size - i] == 1:
out.appendleft(s1[s1_size - i])
else:
assert s1[s1_size - i] == 1, "Incompatible broadcast type %s and %s" % (
s0[s0_size - i],
s1[s1_size - i],
)
out.appendleft(s0[s0_size - i])
if s0_size < s1_size:
for i in range(s0_size + 1, s1_size + 1):
out.appendleft(s1[s1_size - i])
if s1_size < s0_size:
for i in range(s1_size + 1, s0_size + 1):
out.appendleft(s0[s0_size - i])
Copy link
Contributor

Choose a reason for hiding this comment

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

Is it better to use itertools.zip_longest() on the reversed array instead of running three separate loops? Might be helpful to have fill_values set at 1 to avoid handling None values.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

This will require atleast two list inversions (both the inputs as we have to iterate in reverse to verify broadcasting rules). Fill values will only make sense if I use a zip. It is only two loops (one of last two will fail) and merging them to one will give marginal optimization.

return _expr.const(list(out), attr["T"].name)

return _impl


def _broadcast_to():
def _impl(inputs, attr, params, mod):
if isinstance(inputs[1], _expr.Var):
Expand Down Expand Up @@ -2740,6 +2778,7 @@ def _impl(inputs, attr, params, mod):
"BatchToSpaceND": _batch_to_space_nd(),
"BiasAdd": _bias_add(),
"BroadcastTo": _broadcast_to(),
"BroadcastArgs": _broadcast_args(),
"Cast": _cast(),
"Ceil": AttrCvt("ceil"),
"CheckNumerics": _check_numerics(),
Expand Down Expand Up @@ -2833,6 +2872,7 @@ def _impl(inputs, attr, params, mod):
"Round": AttrCvt("round"),
"Rsqrt": _rsqrt(),
"Select": _where(),
"SelectV2": _where(),
comaniac marked this conversation as resolved.
Show resolved Hide resolved
"Selu": _selu(),
"Shape": _shape(),
"Sigmoid": AttrCvt("sigmoid"),
Expand Down Expand Up @@ -3936,7 +3976,6 @@ def _backtrack_construct(self, node_name):
raise ImportError("Unable to import tensorflow which is required {}".format(e))

input_op_name = node_name.split(":")[0].split("^")[-1]

if input_op_name not in self._nodes:
node = self._tf_node_map[input_op_name]
attr = self._parse_attr(node.attr)
Expand Down Expand Up @@ -3997,7 +4036,6 @@ def _backtrack_construct(self, node_name):
inputs[i] = actual_input

op = self._convert_operator(node.op, node.name, inputs, attr)

if isinstance(op, np.ndarray):
self._params[node.name] = tvm.nd.array(op)
op = [
Expand All @@ -4019,7 +4057,6 @@ def _backtrack_construct(self, node_name):
tn = node_name.split(":")
tensor_slot = int(tn[1]) if len(tn) > 1 else 0
return out[tensor_slot]

return out[0]


Expand Down
27 changes: 27 additions & 0 deletions tests/python/frontend/tensorflow/test_forward.py
Original file line number Diff line number Diff line change
Expand Up @@ -3050,6 +3050,33 @@ def test_forward_resize():
_test_resize_nearest_neighbor_dynamic_shape((1, 16, 16, 3), scale=[2, 2])


#######################################################################
# BroadcastArgs
# -----------


def _test_broadcast_args(in_shape_1, in_shape_2):
""" One iteration of broadcast_args"""

shape_1 = np.array(in_shape_1).astype("int32")
shape_2 = np.array(in_shape_2).astype("int32")

with tf.Graph().as_default():
shape_1 = constant_op.constant(shape_1, shape=shape_1.shape, dtype=shape_1.dtype)
shape_2 = constant_op.constant(shape_2, shape=shape_2.shape, dtype=shape_2.dtype)
tf.raw_ops.BroadcastArgs(s0=shape_1, s1=shape_2)

compare_tf_with_tvm(None, "", "BroadcastArgs:0", opt_level=0)


def test_forward_broadcast_args():
""" Resize Bilinear """

_test_broadcast_args((4, 1, 32, 32), [4, 8, 32, 32])
_test_broadcast_args((6, 32, 32, 1), [6, 32, 32, 16])
_test_broadcast_args((32, 32, 16), [6, 32, 32, 16])


#######################################################################
# BroadcastTo
# -----------
Expand Down