Skip to content

div_groups #23

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

Merged
merged 1 commit into from
Aug 26, 2021
Merged

div_groups #23

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
93 changes: 80 additions & 13 deletions Nbs/04_YaResNet.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,7 @@
},
{
"cell_type": "code",
"execution_count": 5,
"execution_count": 1,
"source": [
"#hide\n",
"# from nbdev.showdoc import *\n",
Expand All @@ -36,7 +36,7 @@
},
{
"cell_type": "code",
"execution_count": 1,
"execution_count": 2,
"source": [
"#hide\n",
"from model_constructor.yaresnet import YaResBlock"
Expand All @@ -46,7 +46,7 @@
},
{
"cell_type": "code",
"execution_count": 2,
"execution_count": 3,
"source": [
"#collapse_output\n",
"bl = YaResBlock(1,64,64,sa=True)\n",
Expand Down Expand Up @@ -77,14 +77,14 @@
]
},
"metadata": {},
"execution_count": 2
"execution_count": 3
}
],
"metadata": {}
},
{
"cell_type": "code",
"execution_count": 6,
"execution_count": 4,
"source": [
"#hide\n",
"bs_test = 16\n",
Expand All @@ -106,7 +106,7 @@
},
{
"cell_type": "code",
"execution_count": 7,
"execution_count": 5,
"source": [
"#collapse_output\n",
"bl = YaResBlock(1,64,64,se=True)\n",
Expand All @@ -131,9 +131,9 @@
" (se): SEBlock(\n",
" (squeeze): AdaptiveAvgPool2d(output_size=1)\n",
" (excitation): Sequential(\n",
" (fc_reduce): Linear(in_features=64, out_features=4, bias=False)\n",
" (fc_reduce): Linear(in_features=64, out_features=4, bias=True)\n",
" (se_act): ReLU(inplace=True)\n",
" (fc_expand): Linear(in_features=4, out_features=64, bias=False)\n",
" (fc_expand): Linear(in_features=4, out_features=64, bias=True)\n",
" (sigmoid): Sigmoid()\n",
" )\n",
" )\n",
Expand All @@ -143,14 +143,14 @@
]
},
"metadata": {},
"execution_count": 7
"execution_count": 5
}
],
"metadata": {}
},
{
"cell_type": "code",
"execution_count": 8,
"execution_count": 6,
"source": [
"#hide\n",
"bs_test = 16\n",
Expand All @@ -172,7 +172,7 @@
},
{
"cell_type": "code",
"execution_count": 9,
"execution_count": 7,
"source": [
"#collapse_output\n",
"bl = YaResBlock(4,64,128,stride=2,act_fn=nn.LeakyReLU(), bn_1st=False)\n",
Expand Down Expand Up @@ -210,6 +210,73 @@
]
},
"metadata": {},
"execution_count": 7
}
],
"metadata": {}
},
{
"cell_type": "code",
"execution_count": 8,
"source": [
"#hide\n",
"bs_test = 16\n",
"xb = torch.randn(bs_test, 256, 32, 32)\n",
"y = bl(xb)\n",
"print(y.shape)\n",
"assert y.shape == torch.Size([bs_test, 512, 16, 16]), f\"size\""
],
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"torch.Size([16, 512, 16, 16])\n"
]
}
],
"metadata": {}
},
{
"cell_type": "code",
"execution_count": 9,
"source": [
"#collapse_output\n",
"bl = YaResBlock(4,64,128,stride=2,act_fn=nn.LeakyReLU(), bn_1st=False, groups=4)\n",
"bl"
],
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"YaResBlock(\n",
" (reduce): AvgPool2d(kernel_size=2, stride=2, padding=0)\n",
" (convs): Sequential(\n",
" (conv_0): ConvLayer(\n",
" (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
" (act_fn): LeakyReLU(negative_slope=0.01)\n",
" (bn): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" )\n",
" (conv_1): ConvLayer(\n",
" (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=4, bias=False)\n",
" (act_fn): LeakyReLU(negative_slope=0.01)\n",
" (bn): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" )\n",
" (conv_2): ConvLayer(\n",
" (conv): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
" (bn): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" )\n",
" )\n",
" (idconv): ConvLayer(\n",
" (conv): Conv2d(256, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
" (bn): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" )\n",
" (merge): LeakyReLU(negative_slope=0.01)\n",
")"
]
},
"metadata": {},
"execution_count": 9
}
],
Expand Down Expand Up @@ -242,7 +309,7 @@
"execution_count": 11,
"source": [
"#collapse_output\n",
"bl = YaResBlock(4,64,128,stride=2,act_fn=nn.LeakyReLU(), bn_1st=False, groups=4)\n",
"bl = YaResBlock(4,64,128,stride=2,act_fn=nn.LeakyReLU(), bn_1st=False, div_groups=4)\n",
"bl"
],
"outputs": [
Expand All @@ -259,7 +326,7 @@
" (bn): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" )\n",
" (conv_1): ConvLayer(\n",
" (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=4, bias=False)\n",
" (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=32, bias=False)\n",
" (act_fn): LeakyReLU(negative_slope=0.01)\n",
" (bn): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
" )\n",
Expand Down
18 changes: 9 additions & 9 deletions model_constructor/net.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,11 +29,11 @@ class ResBlock(nn.Module):
def __init__(self, expansion, ni, nh, stride=1,
conv_layer=ConvLayer, act_fn=act_fn, zero_bn=True, bn_1st=True,
pool=nn.AvgPool2d(2, ceil_mode=True), sa=False, sym=False, se=False, se_reduction=16,
groups=1, dw=False):
groups=1, dw=False, div_groups=None):
super().__init__()
nf, ni = nh * expansion, ni * expansion
# if groups != 1:
# groups = int(nh / groups)
if div_groups is not None: # check if grops != 1 and div_groups
groups = int(nh / div_groups)
if expansion == 1:
layers = [("conv_0", conv_layer(ni, nh, 3, stride=stride, act_fn=act_fn, bn_1st=bn_1st,
groups=nh if dw else groups)),
Expand Down Expand Up @@ -68,11 +68,11 @@ class NewResBlock(nn.Module):
def __init__(self, expansion, ni, nh, stride=1,
conv_layer=ConvLayer, act_fn=act_fn, zero_bn=True, bn_1st=True,
pool=nn.AvgPool2d(2, ceil_mode=True), sa=False, sym=False, se=False, se_reduction=16,
groups=1, dw=False):
groups=1, dw=False, div_groups=None):
super().__init__()
nf, ni = nh * expansion, ni * expansion
# if groups != 1:
# groups = int(nh / groups)
if div_groups is not None: # check if grops != 1 and div_groups
groups = int(nh / div_groups)
self.reduce = noop if stride == 1 else pool
if expansion == 1:
layers = [("conv_0", conv_layer(ni, nh, 3, stride=1, act_fn=act_fn, bn_1st=bn_1st,
Expand Down Expand Up @@ -114,7 +114,7 @@ def _make_layer(self, expansion, ni, nf, blocks, stride, sa):
layers = [(f"bl_{i}", self.block(expansion, ni if i == 0 else nf, nf,
stride if i == 0 else 1, sa=sa if i == blocks - 1 else False,
conv_layer=self.conv_layer, act_fn=self.act_fn, pool=self.pool,
zero_bn=self.zero_bn, bn_1st=self.bn_1st, groups=self.groups,
zero_bn=self.zero_bn, bn_1st=self.bn_1st, groups=self.groups, div_groups=self.div_groups,
dw=self.dw, se=self.se))
for i in range(blocks)]
return nn.Sequential(OrderedDict(layers))
Expand Down Expand Up @@ -144,7 +144,7 @@ def __init__(self, name='Net', c_in=3, c_out=1000,
norm=nn.BatchNorm2d,
act_fn=nn.ReLU(inplace=True),
pool=nn.AvgPool2d(2, ceil_mode=True),
expansion=1, groups=1, dw=False,
expansion=1, groups=1, dw=False, div_groups=None,
sa=False, se=False, se_reduction=16,
bn_1st=True,
zero_bn=True,
Expand Down Expand Up @@ -195,7 +195,7 @@ def __call__(self):
def __repr__(self):
return (f"{self.name} constructor\n"
f" c_in: {self.c_in}, c_out: {self.c_out}\n"
f" expansion: {self.expansion}, groups: {self.groups}, dw: {self.dw}\n"
f" expansion: {self.expansion}, groups: {self.groups}, dw: {self.dw}\n, div_groups: {self.div_groups}"
f" sa: {self.sa}, se: {self.se}\n"
f" stem sizes: {self.stem_sizes}, stide on {self.stem_stride_on}\n"
f" body sizes {self._block_sizes}\n"
Expand Down
6 changes: 3 additions & 3 deletions model_constructor/yaresnet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,11 +19,11 @@ class YaResBlock(nn.Module):
def __init__(self, expansion, ni, nh, stride=1,
conv_layer=ConvLayer, act_fn=act_fn, zero_bn=True, bn_1st=True,
pool=nn.AvgPool2d(2, ceil_mode=True), sa=False, sym=False, se=False,
groups=1, dw=False):
groups=1, dw=False, div_groups=None):
super().__init__()
nf, ni = nh * expansion, ni * expansion
# if groups != 1:
# groups = int(nh / groups)
if div_groups is not None: # check if grops != 1 and div_groups
groups = int(nh / div_groups)
self.reduce = noop if stride == 1 else pool
layers = [("conv_0", conv_layer(ni, nh, 3, stride=1, act_fn=act_fn, bn_1st=bn_1st,
groups=nh if dw else groups)),
Expand Down