Add int search space
This commit is contained in:
D-X-Y
@@ -21,149 +21,209 @@ import torch.nn.functional as F
|
||||
|
||||
|
||||
def drop_block_2d(
|
||||
x, drop_prob: float = 0.1, block_size: int = 7, gamma_scale: float = 1.0,
|
||||
with_noise: bool = False, inplace: bool = False, batchwise: bool = False):
|
||||
""" DropBlock. See https://arxiv.org/pdf/1810.12890.pdf
|
||||
x,
|
||||
drop_prob: float = 0.1,
|
||||
block_size: int = 7,
|
||||
gamma_scale: float = 1.0,
|
||||
with_noise: bool = False,
|
||||
inplace: bool = False,
|
||||
batchwise: bool = False,
|
||||
):
|
||||
"""DropBlock. See https://arxiv.org/pdf/1810.12890.pdf
|
||||
|
||||
DropBlock with an experimental gaussian noise option. This layer has been tested on a few training
|
||||
runs with success, but needs further validation and possibly optimization for lower runtime impact.
|
||||
"""
|
||||
B, C, H, W = x.shape
|
||||
total_size = W * H
|
||||
clipped_block_size = min(block_size, min(W, H))
|
||||
# seed_drop_rate, the gamma parameter
|
||||
gamma = gamma_scale * drop_prob * total_size / clipped_block_size ** 2 / (
|
||||
(W - block_size + 1) * (H - block_size + 1))
|
||||
DropBlock with an experimental gaussian noise option. This layer has been tested on a few training
|
||||
runs with success, but needs further validation and possibly optimization for lower runtime impact.
|
||||
"""
|
||||
B, C, H, W = x.shape
|
||||
total_size = W * H
|
||||
clipped_block_size = min(block_size, min(W, H))
|
||||
# seed_drop_rate, the gamma parameter
|
||||
gamma = (
|
||||
gamma_scale
|
||||
* drop_prob
|
||||
* total_size
|
||||
/ clipped_block_size ** 2
|
||||
/ ((W - block_size + 1) * (H - block_size + 1))
|
||||
)
|
||||
|
||||
# Forces the block to be inside the feature map.
|
||||
w_i, h_i = torch.meshgrid(torch.arange(W).to(x.device), torch.arange(H).to(x.device))
|
||||
valid_block = ((w_i >= clipped_block_size // 2) & (w_i < W - (clipped_block_size - 1) // 2)) & \
|
||||
((h_i >= clipped_block_size // 2) & (h_i < H - (clipped_block_size - 1) // 2))
|
||||
valid_block = torch.reshape(valid_block, (1, 1, H, W)).to(dtype=x.dtype)
|
||||
# Forces the block to be inside the feature map.
|
||||
w_i, h_i = torch.meshgrid(
|
||||
torch.arange(W).to(x.device), torch.arange(H).to(x.device)
|
||||
)
|
||||
valid_block = (
|
||||
(w_i >= clipped_block_size // 2) & (w_i < W - (clipped_block_size - 1) // 2)
|
||||
) & ((h_i >= clipped_block_size // 2) & (h_i < H - (clipped_block_size - 1) // 2))
|
||||
valid_block = torch.reshape(valid_block, (1, 1, H, W)).to(dtype=x.dtype)
|
||||
|
||||
if batchwise:
|
||||
# one mask for whole batch, quite a bit faster
|
||||
uniform_noise = torch.rand((1, C, H, W), dtype=x.dtype, device=x.device)
|
||||
else:
|
||||
uniform_noise = torch.rand_like(x)
|
||||
block_mask = ((2 - gamma - valid_block + uniform_noise) >= 1).to(dtype=x.dtype)
|
||||
block_mask = -F.max_pool2d(
|
||||
-block_mask,
|
||||
kernel_size=clipped_block_size, # block_size,
|
||||
stride=1,
|
||||
padding=clipped_block_size // 2)
|
||||
|
||||
if with_noise:
|
||||
normal_noise = torch.randn((1, C, H, W), dtype=x.dtype, device=x.device) if batchwise else torch.randn_like(x)
|
||||
if inplace:
|
||||
x.mul_(block_mask).add_(normal_noise * (1 - block_mask))
|
||||
if batchwise:
|
||||
# one mask for whole batch, quite a bit faster
|
||||
uniform_noise = torch.rand((1, C, H, W), dtype=x.dtype, device=x.device)
|
||||
else:
|
||||
x = x * block_mask + normal_noise * (1 - block_mask)
|
||||
else:
|
||||
normalize_scale = (block_mask.numel() / block_mask.to(dtype=torch.float32).sum().add(1e-7)).to(x.dtype)
|
||||
if inplace:
|
||||
x.mul_(block_mask * normalize_scale)
|
||||
uniform_noise = torch.rand_like(x)
|
||||
block_mask = ((2 - gamma - valid_block + uniform_noise) >= 1).to(dtype=x.dtype)
|
||||
block_mask = -F.max_pool2d(
|
||||
-block_mask,
|
||||
kernel_size=clipped_block_size, # block_size,
|
||||
stride=1,
|
||||
padding=clipped_block_size // 2,
|
||||
)
|
||||
|
||||
if with_noise:
|
||||
normal_noise = (
|
||||
torch.randn((1, C, H, W), dtype=x.dtype, device=x.device)
|
||||
if batchwise
|
||||
else torch.randn_like(x)
|
||||
)
|
||||
if inplace:
|
||||
x.mul_(block_mask).add_(normal_noise * (1 - block_mask))
|
||||
else:
|
||||
x = x * block_mask + normal_noise * (1 - block_mask)
|
||||
else:
|
||||
x = x * block_mask * normalize_scale
|
||||
return x
|
||||
normalize_scale = (
|
||||
block_mask.numel() / block_mask.to(dtype=torch.float32).sum().add(1e-7)
|
||||
).to(x.dtype)
|
||||
if inplace:
|
||||
x.mul_(block_mask * normalize_scale)
|
||||
else:
|
||||
x = x * block_mask * normalize_scale
|
||||
return x
|
||||
|
||||
|
||||
def drop_block_fast_2d(
|
||||
x: torch.Tensor, drop_prob: float = 0.1, block_size: int = 7,
|
||||
gamma_scale: float = 1.0, with_noise: bool = False, inplace: bool = False, batchwise: bool = False):
|
||||
""" DropBlock. See https://arxiv.org/pdf/1810.12890.pdf
|
||||
x: torch.Tensor,
|
||||
drop_prob: float = 0.1,
|
||||
block_size: int = 7,
|
||||
gamma_scale: float = 1.0,
|
||||
with_noise: bool = False,
|
||||
inplace: bool = False,
|
||||
batchwise: bool = False,
|
||||
):
|
||||
"""DropBlock. See https://arxiv.org/pdf/1810.12890.pdf
|
||||
|
||||
DropBlock with an experimental gaussian noise option. Simplied from above without concern for valid
|
||||
block mask at edges.
|
||||
"""
|
||||
B, C, H, W = x.shape
|
||||
total_size = W * H
|
||||
clipped_block_size = min(block_size, min(W, H))
|
||||
gamma = gamma_scale * drop_prob * total_size / clipped_block_size ** 2 / (
|
||||
(W - block_size + 1) * (H - block_size + 1))
|
||||
DropBlock with an experimental gaussian noise option. Simplied from above without concern for valid
|
||||
block mask at edges.
|
||||
"""
|
||||
B, C, H, W = x.shape
|
||||
total_size = W * H
|
||||
clipped_block_size = min(block_size, min(W, H))
|
||||
gamma = (
|
||||
gamma_scale
|
||||
* drop_prob
|
||||
* total_size
|
||||
/ clipped_block_size ** 2
|
||||
/ ((W - block_size + 1) * (H - block_size + 1))
|
||||
)
|
||||
|
||||
if batchwise:
|
||||
# one mask for whole batch, quite a bit faster
|
||||
block_mask = torch.rand((1, C, H, W), dtype=x.dtype, device=x.device) < gamma
|
||||
else:
|
||||
# mask per batch element
|
||||
block_mask = torch.rand_like(x) < gamma
|
||||
block_mask = F.max_pool2d(
|
||||
block_mask.to(x.dtype), kernel_size=clipped_block_size, stride=1, padding=clipped_block_size // 2)
|
||||
|
||||
if with_noise:
|
||||
normal_noise = torch.randn((1, C, H, W), dtype=x.dtype, device=x.device) if batchwise else torch.randn_like(x)
|
||||
if inplace:
|
||||
x.mul_(1. - block_mask).add_(normal_noise * block_mask)
|
||||
if batchwise:
|
||||
# one mask for whole batch, quite a bit faster
|
||||
block_mask = torch.rand((1, C, H, W), dtype=x.dtype, device=x.device) < gamma
|
||||
else:
|
||||
x = x * (1. - block_mask) + normal_noise * block_mask
|
||||
else:
|
||||
block_mask = 1 - block_mask
|
||||
normalize_scale = (block_mask.numel() / block_mask.to(dtype=torch.float32).sum().add(1e-7)).to(dtype=x.dtype)
|
||||
if inplace:
|
||||
x.mul_(block_mask * normalize_scale)
|
||||
# mask per batch element
|
||||
block_mask = torch.rand_like(x) < gamma
|
||||
block_mask = F.max_pool2d(
|
||||
block_mask.to(x.dtype),
|
||||
kernel_size=clipped_block_size,
|
||||
stride=1,
|
||||
padding=clipped_block_size // 2,
|
||||
)
|
||||
|
||||
if with_noise:
|
||||
normal_noise = (
|
||||
torch.randn((1, C, H, W), dtype=x.dtype, device=x.device)
|
||||
if batchwise
|
||||
else torch.randn_like(x)
|
||||
)
|
||||
if inplace:
|
||||
x.mul_(1.0 - block_mask).add_(normal_noise * block_mask)
|
||||
else:
|
||||
x = x * (1.0 - block_mask) + normal_noise * block_mask
|
||||
else:
|
||||
x = x * block_mask * normalize_scale
|
||||
return x
|
||||
block_mask = 1 - block_mask
|
||||
normalize_scale = (
|
||||
block_mask.numel() / block_mask.to(dtype=torch.float32).sum().add(1e-7)
|
||||
).to(dtype=x.dtype)
|
||||
if inplace:
|
||||
x.mul_(block_mask * normalize_scale)
|
||||
else:
|
||||
x = x * block_mask * normalize_scale
|
||||
return x
|
||||
|
||||
|
||||
class DropBlock2d(nn.Module):
|
||||
""" DropBlock. See https://arxiv.org/pdf/1810.12890.pdf
|
||||
"""
|
||||
def __init__(self,
|
||||
drop_prob=0.1,
|
||||
block_size=7,
|
||||
gamma_scale=1.0,
|
||||
with_noise=False,
|
||||
inplace=False,
|
||||
batchwise=False,
|
||||
fast=True):
|
||||
super(DropBlock2d, self).__init__()
|
||||
self.drop_prob = drop_prob
|
||||
self.gamma_scale = gamma_scale
|
||||
self.block_size = block_size
|
||||
self.with_noise = with_noise
|
||||
self.inplace = inplace
|
||||
self.batchwise = batchwise
|
||||
self.fast = fast # FIXME finish comparisons of fast vs not
|
||||
"""DropBlock. See https://arxiv.org/pdf/1810.12890.pdf"""
|
||||
|
||||
def forward(self, x):
|
||||
if not self.training or not self.drop_prob:
|
||||
return x
|
||||
if self.fast:
|
||||
return drop_block_fast_2d(
|
||||
x, self.drop_prob, self.block_size, self.gamma_scale, self.with_noise, self.inplace, self.batchwise)
|
||||
else:
|
||||
return drop_block_2d(
|
||||
x, self.drop_prob, self.block_size, self.gamma_scale, self.with_noise, self.inplace, self.batchwise)
|
||||
def __init__(
|
||||
self,
|
||||
drop_prob=0.1,
|
||||
block_size=7,
|
||||
gamma_scale=1.0,
|
||||
with_noise=False,
|
||||
inplace=False,
|
||||
batchwise=False,
|
||||
fast=True,
|
||||
):
|
||||
super(DropBlock2d, self).__init__()
|
||||
self.drop_prob = drop_prob
|
||||
self.gamma_scale = gamma_scale
|
||||
self.block_size = block_size
|
||||
self.with_noise = with_noise
|
||||
self.inplace = inplace
|
||||
self.batchwise = batchwise
|
||||
self.fast = fast # FIXME finish comparisons of fast vs not
|
||||
|
||||
def forward(self, x):
|
||||
if not self.training or not self.drop_prob:
|
||||
return x
|
||||
if self.fast:
|
||||
return drop_block_fast_2d(
|
||||
x,
|
||||
self.drop_prob,
|
||||
self.block_size,
|
||||
self.gamma_scale,
|
||||
self.with_noise,
|
||||
self.inplace,
|
||||
self.batchwise,
|
||||
)
|
||||
else:
|
||||
return drop_block_2d(
|
||||
x,
|
||||
self.drop_prob,
|
||||
self.block_size,
|
||||
self.gamma_scale,
|
||||
self.with_noise,
|
||||
self.inplace,
|
||||
self.batchwise,
|
||||
)
|
||||
|
||||
|
||||
def drop_path(x, drop_prob: float = 0., training: bool = False):
|
||||
"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).
|
||||
def drop_path(x, drop_prob: float = 0.0, training: bool = False):
|
||||
"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).
|
||||
|
||||
This is the same as the DropConnect impl I created for EfficientNet, etc networks, however,
|
||||
the original name is misleading as 'Drop Connect' is a different form of dropout in a separate paper...
|
||||
See discussion: https://github.com/tensorflow/tpu/issues/494#issuecomment-532968956 ... I've opted for
|
||||
changing the layer and argument names to 'drop path' rather than mix DropConnect as a layer name and use
|
||||
'survival rate' as the argument.
|
||||
This is the same as the DropConnect impl I created for EfficientNet, etc networks, however,
|
||||
the original name is misleading as 'Drop Connect' is a different form of dropout in a separate paper...
|
||||
See discussion: https://github.com/tensorflow/tpu/issues/494#issuecomment-532968956 ... I've opted for
|
||||
changing the layer and argument names to 'drop path' rather than mix DropConnect as a layer name and use
|
||||
'survival rate' as the argument.
|
||||
|
||||
"""
|
||||
if drop_prob == 0. or not training:
|
||||
return x
|
||||
keep_prob = 1 - drop_prob
|
||||
shape = (x.shape[0],) + (1,) * (x.ndim - 1) # work with diff dim tensors, not just 2D ConvNets
|
||||
random_tensor = keep_prob + torch.rand(shape, dtype=x.dtype, device=x.device)
|
||||
random_tensor.floor_() # binarize
|
||||
output = x.div(keep_prob) * random_tensor
|
||||
return output
|
||||
"""
|
||||
if drop_prob == 0.0 or not training:
|
||||
return x
|
||||
keep_prob = 1 - drop_prob
|
||||
shape = (x.shape[0],) + (1,) * (
|
||||
x.ndim - 1
|
||||
) # work with diff dim tensors, not just 2D ConvNets
|
||||
random_tensor = keep_prob + torch.rand(shape, dtype=x.dtype, device=x.device)
|
||||
random_tensor.floor_() # binarize
|
||||
output = x.div(keep_prob) * random_tensor
|
||||
return output
|
||||
|
||||
|
||||
class DropPath(nn.Module):
|
||||
"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).
|
||||
"""
|
||||
def __init__(self, drop_prob=None):
|
||||
super(DropPath, self).__init__()
|
||||
self.drop_prob = drop_prob
|
||||
"""Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks)."""
|
||||
|
||||
def forward(self, x):
|
||||
return drop_path(x, self.drop_prob, self.training)
|
||||
def __init__(self, drop_prob=None):
|
||||
super(DropPath, self).__init__()
|
||||
self.drop_prob = drop_prob
|
||||
|
||||
def forward(self, x):
|
||||
return drop_path(x, self.drop_prob, self.training)
|
||||
|
||||
@@ -1,24 +1,29 @@
|
||||
import torch.nn as nn
|
||||
from typing import Optional
|
||||
|
||||
class MLP(nn.Module):
|
||||
# MLP: FC -> Activation -> Drop -> FC -> Drop
|
||||
def __init__(self, in_features, hidden_features: Optional[int] = None,
|
||||
out_features: Optional[int] = None,
|
||||
act_layer=nn.GELU,
|
||||
drop: Optional[float] = None):
|
||||
super(MLP, self).__init__()
|
||||
out_features = out_features or in_features
|
||||
hidden_features = hidden_features or in_features
|
||||
self.fc1 = nn.Linear(in_features, hidden_features)
|
||||
self.act = act_layer()
|
||||
self.fc2 = nn.Linear(hidden_features, out_features)
|
||||
self.drop = nn.Dropout(drop or 0)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.fc1(x)
|
||||
x = self.act(x)
|
||||
x = self.drop(x)
|
||||
x = self.fc2(x)
|
||||
x = self.drop(x)
|
||||
return x
|
||||
class MLP(nn.Module):
|
||||
# MLP: FC -> Activation -> Drop -> FC -> Drop
|
||||
def __init__(
|
||||
self,
|
||||
in_features,
|
||||
hidden_features: Optional[int] = None,
|
||||
out_features: Optional[int] = None,
|
||||
act_layer=nn.GELU,
|
||||
drop: Optional[float] = None,
|
||||
):
|
||||
super(MLP, self).__init__()
|
||||
out_features = out_features or in_features
|
||||
hidden_features = hidden_features or in_features
|
||||
self.fc1 = nn.Linear(in_features, hidden_features)
|
||||
self.act = act_layer()
|
||||
self.fc2 = nn.Linear(hidden_features, out_features)
|
||||
self.drop = nn.Dropout(drop or 0)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.fc1(x)
|
||||
x = self.act(x)
|
||||
x = self.drop(x)
|
||||
x = self.fc2(x)
|
||||
x = self.drop(x)
|
||||
return x
|
||||
|
||||
@@ -5,31 +5,31 @@ import torch
|
||||
import torch.nn as nn
|
||||
import math
|
||||
|
||||
class PositionalEncoder(nn.Module):
|
||||
# Attention Is All You Need: https://arxiv.org/pdf/1706.03762.pdf
|
||||
# https://github.com/pytorch/examples/blob/master/word_language_model/model.py#L65
|
||||
|
||||
def __init__(self, d_model, max_seq_len, dropout=0.1):
|
||||
super(PositionalEncoder, self).__init__()
|
||||
self.d_model = d_model
|
||||
# create constant 'pe' matrix with values dependant on
|
||||
# pos and i
|
||||
pe = torch.zeros(max_seq_len, d_model)
|
||||
for pos in range(max_seq_len):
|
||||
for i in range(0, d_model):
|
||||
div = 10000 ** ((i // 2) * 2 / d_model)
|
||||
value = pos / div
|
||||
if i % 2 == 0:
|
||||
pe[pos, i] = math.sin(value)
|
||||
else:
|
||||
pe[pos, i] = math.cos(value)
|
||||
pe = pe.unsqueeze(0)
|
||||
self.dropout = nn.Dropout(p=dropout)
|
||||
self.register_buffer('pe', pe)
|
||||
|
||||
|
||||
def forward(self, x):
|
||||
batch, seq, fdim = x.shape[:3]
|
||||
embeddings = self.pe[:, :seq, :fdim]
|
||||
outs = self.dropout(x + embeddings)
|
||||
return outs
|
||||
class PositionalEncoder(nn.Module):
|
||||
# Attention Is All You Need: https://arxiv.org/pdf/1706.03762.pdf
|
||||
# https://github.com/pytorch/examples/blob/master/word_language_model/model.py#L65
|
||||
|
||||
def __init__(self, d_model, max_seq_len, dropout=0.1):
|
||||
super(PositionalEncoder, self).__init__()
|
||||
self.d_model = d_model
|
||||
# create constant 'pe' matrix with values dependant on
|
||||
# pos and i
|
||||
pe = torch.zeros(max_seq_len, d_model)
|
||||
for pos in range(max_seq_len):
|
||||
for i in range(0, d_model):
|
||||
div = 10000 ** ((i // 2) * 2 / d_model)
|
||||
value = pos / div
|
||||
if i % 2 == 0:
|
||||
pe[pos, i] = math.sin(value)
|
||||
else:
|
||||
pe[pos, i] = math.cos(value)
|
||||
pe = pe.unsqueeze(0)
|
||||
self.dropout = nn.Dropout(p=dropout)
|
||||
self.register_buffer("pe", pe)
|
||||
|
||||
def forward(self, x):
|
||||
batch, seq, fdim = x.shape[:3]
|
||||
embeddings = self.pe[:, :seq, :fdim]
|
||||
outs = self.dropout(x + embeddings)
|
||||
return outs
|
||||
|
||||
@@ -2,24 +2,22 @@ import torch.nn as nn
|
||||
from torch.nn.parameter import Parameter
|
||||
from typing import Optional
|
||||
|
||||
from layers.super_module import SuperModule
|
||||
from layers.super_module import SuperModule
|
||||
|
||||
class Linear(nn.Module):
|
||||
"""Applies a linear transformation to the incoming data: :math:`y = xA^T + b`
|
||||
"""
|
||||
__constants__ = ['in_features', 'out_features']
|
||||
in_features: int
|
||||
out_features: int
|
||||
weight: Tensor
|
||||
|
||||
class SuperLinear(SuperModule):
|
||||
"""Applies a linear transformation to the incoming data: :math:`y = xA^T + b`"""
|
||||
|
||||
def __init__(self, in_features: int, out_features: int, bias: bool = True) -> None:
|
||||
super(Linear, self).__init__()
|
||||
super(SuperLinear, self).__init__()
|
||||
self.in_features = in_features
|
||||
self.out_features = out_features
|
||||
self.weight = Parameter(torch.Tensor(out_features, in_features))
|
||||
if bias:
|
||||
self.bias = Parameter(torch.Tensor(out_features))
|
||||
else:
|
||||
self.register_parameter('bias', None)
|
||||
self.register_parameter("bias", None)
|
||||
self.reset_parameters()
|
||||
|
||||
def reset_parameters(self) -> None:
|
||||
@@ -33,28 +31,33 @@ class Linear(nn.Module):
|
||||
return F.linear(input, self.weight, self.bias)
|
||||
|
||||
def extra_repr(self) -> str:
|
||||
return 'in_features={}, out_features={}, bias={}'.format(
|
||||
return "in_features={:}, out_features={:}, bias={:}".format(
|
||||
self.in_features, self.out_features, self.bias is not None
|
||||
)
|
||||
|
||||
class SuperMLP(nn.Module):
|
||||
# MLP: FC -> Activation -> Drop -> FC -> Drop
|
||||
def __init__(self, in_features, hidden_features: Optional[int] = None,
|
||||
out_features: Optional[int] = None,
|
||||
act_layer=nn.GELU,
|
||||
drop: Optional[float] = None):
|
||||
super(MLP, self).__init__()
|
||||
out_features = out_features or in_features
|
||||
hidden_features = hidden_features or in_features
|
||||
self.fc1 = nn.Linear(in_features, hidden_features)
|
||||
self.act = act_layer()
|
||||
self.fc2 = nn.Linear(hidden_features, out_features)
|
||||
self.drop = nn.Dropout(drop or 0)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.fc1(x)
|
||||
x = self.act(x)
|
||||
x = self.drop(x)
|
||||
x = self.fc2(x)
|
||||
x = self.drop(x)
|
||||
return x
|
||||
class SuperMLP(nn.Module):
|
||||
# MLP: FC -> Activation -> Drop -> FC -> Drop
|
||||
def __init__(
|
||||
self,
|
||||
in_features,
|
||||
hidden_features: Optional[int] = None,
|
||||
out_features: Optional[int] = None,
|
||||
act_layer=nn.GELU,
|
||||
drop: Optional[float] = None,
|
||||
):
|
||||
super(MLP, self).__init__()
|
||||
out_features = out_features or in_features
|
||||
hidden_features = hidden_features or in_features
|
||||
self.fc1 = nn.Linear(in_features, hidden_features)
|
||||
self.act = act_layer()
|
||||
self.fc2 = nn.Linear(hidden_features, out_features)
|
||||
self.drop = nn.Dropout(drop or 0)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.fc1(x)
|
||||
x = self.act(x)
|
||||
x = self.drop(x)
|
||||
x = self.fc2(x)
|
||||
x = self.drop(x)
|
||||
return x
|
||||
|
||||
@@ -1,7 +1,17 @@
|
||||
#####################################################
|
||||
# Copyright (c) Xuanyi Dong [GitHub D-X-Y], 2021.01 #
|
||||
#####################################################
|
||||
|
||||
import abc
|
||||
import torch.nn as nn
|
||||
|
||||
class SuperModule(nn.Module):
|
||||
def __init__(self):
|
||||
super(SuperModule, self).__init__()
|
||||
|
||||
|
||||
class SuperModule(abc.ABCMeta, nn.Module):
|
||||
"""This class equips the nn.Module class with the ability to apply AutoDL."""
|
||||
|
||||
def __init__(self):
|
||||
super(SuperModule, self).__init__()
|
||||
|
||||
@abc.abstractmethod
|
||||
def abstract_search_space(self):
|
||||
raise NotImplementedError
|
||||
|
||||
@@ -5,57 +5,59 @@ import warnings
|
||||
|
||||
|
||||
def _no_grad_trunc_normal_(tensor, mean, std, a, b):
|
||||
# Cut & paste from PyTorch official master until it's in a few official releases - RW
|
||||
# Method based on https://people.sc.fsu.edu/~jburkardt/presentations/truncated_normal.pdf
|
||||
def norm_cdf(x):
|
||||
# Computes standard normal cumulative distribution function
|
||||
return (1. + math.erf(x / math.sqrt(2.))) / 2.
|
||||
# Cut & paste from PyTorch official master until it's in a few official releases - RW
|
||||
# Method based on https://people.sc.fsu.edu/~jburkardt/presentations/truncated_normal.pdf
|
||||
def norm_cdf(x):
|
||||
# Computes standard normal cumulative distribution function
|
||||
return (1.0 + math.erf(x / math.sqrt(2.0))) / 2.0
|
||||
|
||||
if (mean < a - 2 * std) or (mean > b + 2 * std):
|
||||
warnings.warn("mean is more than 2 std from [a, b] in nn.init.trunc_normal_. "
|
||||
"The distribution of values may be incorrect.",
|
||||
stacklevel=2)
|
||||
if (mean < a - 2 * std) or (mean > b + 2 * std):
|
||||
warnings.warn(
|
||||
"mean is more than 2 std from [a, b] in nn.init.trunc_normal_. "
|
||||
"The distribution of values may be incorrect.",
|
||||
stacklevel=2,
|
||||
)
|
||||
|
||||
with torch.no_grad():
|
||||
# Values are generated by using a truncated uniform distribution and
|
||||
# then using the inverse CDF for the normal distribution.
|
||||
# Get upper and lower cdf values
|
||||
l = norm_cdf((a - mean) / std)
|
||||
u = norm_cdf((b - mean) / std)
|
||||
with torch.no_grad():
|
||||
# Values are generated by using a truncated uniform distribution and
|
||||
# then using the inverse CDF for the normal distribution.
|
||||
# Get upper and lower cdf values
|
||||
l = norm_cdf((a - mean) / std)
|
||||
u = norm_cdf((b - mean) / std)
|
||||
|
||||
# Uniformly fill tensor with values from [l, u], then translate to
|
||||
# [2l-1, 2u-1].
|
||||
tensor.uniform_(2 * l - 1, 2 * u - 1)
|
||||
# Uniformly fill tensor with values from [l, u], then translate to
|
||||
# [2l-1, 2u-1].
|
||||
tensor.uniform_(2 * l - 1, 2 * u - 1)
|
||||
|
||||
# Use inverse cdf transform for normal distribution to get truncated
|
||||
# standard normal
|
||||
tensor.erfinv_()
|
||||
# Use inverse cdf transform for normal distribution to get truncated
|
||||
# standard normal
|
||||
tensor.erfinv_()
|
||||
|
||||
# Transform to proper mean, std
|
||||
tensor.mul_(std * math.sqrt(2.))
|
||||
tensor.add_(mean)
|
||||
# Transform to proper mean, std
|
||||
tensor.mul_(std * math.sqrt(2.0))
|
||||
tensor.add_(mean)
|
||||
|
||||
# Clamp to ensure it's in the proper range
|
||||
tensor.clamp_(min=a, max=b)
|
||||
return tensor
|
||||
# Clamp to ensure it's in the proper range
|
||||
tensor.clamp_(min=a, max=b)
|
||||
return tensor
|
||||
|
||||
|
||||
def trunc_normal_(tensor, mean=0., std=1., a=-2., b=2.):
|
||||
# type: (Tensor, float, float, float, float) -> Tensor
|
||||
r"""Fills the input Tensor with values drawn from a truncated
|
||||
normal distribution. The values are effectively drawn from the
|
||||
normal distribution :math:`\mathcal{N}(\text{mean}, \text{std}^2)`
|
||||
with values outside :math:`[a, b]` redrawn until they are within
|
||||
the bounds. The method used for generating the random values works
|
||||
best when :math:`a \leq \text{mean} \leq b`.
|
||||
Args:
|
||||
tensor: an n-dimensional `torch.Tensor`
|
||||
mean: the mean of the normal distribution
|
||||
std: the standard deviation of the normal distribution
|
||||
a: the minimum cutoff value
|
||||
b: the maximum cutoff value
|
||||
Examples:
|
||||
>>> w = torch.empty(3, 5)
|
||||
>>> nn.init.trunc_normal_(w)
|
||||
"""
|
||||
return _no_grad_trunc_normal_(tensor, mean, std, a, b)
|
||||
def trunc_normal_(tensor, mean=0.0, std=1.0, a=-2.0, b=2.0):
|
||||
# type: (Tensor, float, float, float, float) -> Tensor
|
||||
r"""Fills the input Tensor with values drawn from a truncated
|
||||
normal distribution. The values are effectively drawn from the
|
||||
normal distribution :math:`\mathcal{N}(\text{mean}, \text{std}^2)`
|
||||
with values outside :math:`[a, b]` redrawn until they are within
|
||||
the bounds. The method used for generating the random values works
|
||||
best when :math:`a \leq \text{mean} \leq b`.
|
||||
Args:
|
||||
tensor: an n-dimensional `torch.Tensor`
|
||||
mean: the mean of the normal distribution
|
||||
std: the standard deviation of the normal distribution
|
||||
a: the minimum cutoff value
|
||||
b: the maximum cutoff value
|
||||
Examples:
|
||||
>>> w = torch.empty(3, 5)
|
||||
>>> nn.init.trunc_normal_(w)
|
||||
"""
|
||||
return _no_grad_trunc_normal_(tensor, mean, std, a, b)
|
||||
|
||||
Reference in New Issue
Block a user