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

[VTA] YoloV3 Support #4887

Merged
merged 8 commits into from
Feb 26, 2020
Merged

[VTA] YoloV3 Support #4887

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
4b49f34
[VTA] YoloV3 Support
huajsj Feb 14, 2020
93cee67
Fix small(0, or 1 heigh/width) detect frame issue.
huajsj Feb 17, 2020
9216179
add yolov3-tiny turtorial
huajsj Feb 22, 2020
db8e0f0
add os import
huajsj Feb 22, 2020
8d3d071
address review comments.
huajsj Feb 25, 2020
9bb445b
rename tutorial file with a short name.
huajsj Feb 26, 2020
545e4d8
rename deploy_vision_on_vta.py into deploy_classification.py.
huajsj Feb 26, 2020
44ed03a
address review comment, fix plint eror in deploy_detection.py
huajsj Feb 26, 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
114 changes: 106 additions & 8 deletions vta/python/vta/top/graphpack.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,8 @@ def run_opt_pass(expr, opt_pass):
return entry if isinstance(expr, relay.Function) else entry.body

def _to_shape(shape):
""" convert shape into tuple.
"""
return tuple(int(sh) for sh in shape)

def _pack_batch_channel(data, dshape, bfactor, cfactor):
Expand All @@ -55,6 +57,46 @@ def _unpack_batch_channel(data, old_shape):
return data


def _const_shape_match(data, dshape, cfactor_out):
huajsj marked this conversation as resolved.
Show resolved Hide resolved
""" Pad the constant if the shape[0] not divisible by cfactor_out.
"""
pad_width = int(dshape[0]) % cfactor_out
if pad_width != 0:
pad_width = cfactor_out -pad_width
data = op.nn.pad(data, [[0, pad_width], [0, 0], [0, 0]])
dshape = tuple([dshape[0] + pad_width, dshape[1], dshape[2]])
return data, dshape

def _weight_shape_match(data, dshape, channels, cfactor_out, transpose=False):
""" Pad the weight if the shape[0] not divisible by cfactor_out.
"""
pad_width = int(dshape[0]) % cfactor_out
huajsj marked this conversation as resolved.
Show resolved Hide resolved
channels_pad = int(channels) % cfactor_out
if pad_width != 0:
pad_width = cfactor_out - pad_width
data = op.nn.pad(data, [[0, pad_width], [0, 0], [0, 0], [0, 0]])
dshape = tuple([dshape[0] + pad_width, dshape[1], dshape[2], dshape[3]])

if channels_pad != 0:
channels = channels + (cfactor_out - channels_pad)

return data, dshape, channels

def _weight_shape_match_transpose(data, dshape, channels, cfactor_out):
""" Pad the weight if the shape[1] not divisible by cfactor_out.
huajsj marked this conversation as resolved.
Show resolved Hide resolved
"""
pad_width = int(dshape[1]) % cfactor_out
channels_pad = int(channels) % cfactor_out
if pad_width != 0:
pad_width = cfactor_out - pad_width
data = op.nn.pad(data, [[0, 0], [0, pad_width], [0, 0], [0, 0]])
dshape = tuple(dshape[0], [dshape[1] + pad_width, dshape[2], dshape[3]])

if channels_pad != 0:
channels = channels + (cfactor_out - channels_pad)

return data, dshape, channels

def _pack_weight(data, dshape, cfactor):
"""Pack the weight into packed format.
"""
Expand Down Expand Up @@ -106,10 +148,19 @@ def _pack_const(data, dshape, dtype, bfactor, cfactor):
return data


def _get_shape(node):
"""Get the shape of a node.
def _get_tensor_shape(node):
"""Get node shape.
"""
if isinstance(node.checked_type, relay.ty.TensorType):
return _to_shape(node.checked_type.shape)
return []

def _get_tensor_type(node):
"""Get node type.
"""
return _to_shape(node.checked_type.shape)
if isinstance(node.checked_type, relay.ty.TensorType):
return node.checked_type.dtype
return "float32"

def _operator_idx_inc(expr, count_meta, operator_current_idx):
"""Increase operator index
Expand All @@ -136,14 +187,17 @@ def __init__(self, bfactor, cfactor, weight_bits):
self.add = op.op.get("add")
self.multiply = op.op.get("multiply")
self.bias_add = op.op.get("nn.bias_add")
self.pad = op.op.get("nn.pad")
self.upsampling = op.op.get("nn.upsampling")
self.reshape = op.op.get("reshape")
self.number_of_conv2d = 0
super().__init__()

def visit_call(self, call):
""" Visit the children. """
# First visit the children.
oshape = _get_shape(call)
odtype = call.checked_type.dtype
oshape = _get_tensor_shape(call)
odtype = _get_tensor_type(call)
input_types = [arg.checked_type for arg in call.args]
args = [self.visit(arg) for arg in call.args]

Expand All @@ -156,7 +210,7 @@ def visit_call(self, call):
if self.start_pack:
self.start_pack = False
data = args[0]
data_shape = _get_shape(call.args[0])
data_shape = _get_tensor_shape(call.args[0])
return _unpack_batch_channel(data, data_shape)
if self.start_pack:
# Operator cases
Expand All @@ -169,19 +223,25 @@ def visit_call(self, call):
data, weight = args
data_shape = _to_shape(input_types[0].shape)
kernel_shape = _to_shape(input_types[1].shape)
channels = call.attrs.channels
weight, kernel_shape, channels = _weight_shape_match(weight,
kernel_shape,
channels,
self.cfactor)
kernel = _pack_weight(weight, kernel_shape, self.cfactor)
# insert bit packing when necessary
if w_lanes != 1:
assert 8 % w_lanes == 0
kernel = op.bitpack(kernel, lanes=w_lanes)

conv2d = op.nn.conv2d(
data,
kernel,
strides=call.attrs.strides,
padding=call.attrs.padding,
dilation=call.attrs.dilation,
groups=call.attrs.groups,
channels=call.attrs.channels,
channels=channels,
kernel_size=call.attrs.kernel_size,
data_layout=data_layout,
kernel_layout=kernel_layout,
Expand All @@ -198,6 +258,11 @@ def visit_call(self, call):
data, weight = args
data_shape = _to_shape(input_types[0].shape)
kernel_shape = _to_shape(input_types[1].shape)
channels = call.attrs.channels
weight, kernel_shape, channels = _weight_shape_match_transpose(weight,
kernel_shape,
channels,
self.cfactor)
kernel = _pack_weight_conv2d_transpose(weight, kernel_shape, self.cfactor)
conv2d = op.nn.conv2d_transpose(
data,
Expand All @@ -218,8 +283,11 @@ def visit_call(self, call):
pass
elif call.op == self.add and len(input_types[1].shape) == 3:
data, const = args
const, input_shape = _const_shape_match(const,
input_types[1].shape,
self.cfactor)
const = _pack_const(const,
_to_shape(input_types[1].shape),
_to_shape(input_shape),
input_types[1].dtype,
self.bfactor,
self.cfactor)
Expand Down Expand Up @@ -247,6 +315,36 @@ def visit_call(self, call):
input_types[0].dtype == 'int32':
cast = relay.Call(op.op.get('cast'), [args[0]], call.attrs)
return relay.Call(op.op.get('copy'), [cast])
elif call.op == self.pad:
pad_width = call.attrs.pad_width
if len(pad_width) == 6:
pass
elif len(pad_width) == 4:
data, = args
new_pad_width = []
new_pad_width.extend(pad_width)
for _ in range(2):
new_pad_width.append([0, 0])
return op.nn.pad(data,
pad_value=call.attrs.pad_value,
pad_width=new_pad_width)
elif call.op == self.upsampling:
data, = args
scale_h = call.attrs.scale_h
scale_w = call.attrs.scale_w
data_layout = "NCHW%dn%dc" % (self.bfactor, self.cfactor)
method = call.attrs.method
align_corners = call.attrs.align_corners
return op.nn.upsampling(data,
scale_h,
scale_w,
data_layout,
method,
align_corners)
elif call.op == self.reshape and len(input_types[0].shape) == 4:
data, = args
data = op.transpose(data, axes=(0, 4, 1, 5, 2, 3))
return op.reshape(data, input_types[0].shape)

return relay.Call(
self.visit(call.op),
Expand Down
Loading