Enable vqvae to use a switched_conv variant
This commit is contained in:
James Betker
parent
41b7d50944
commit
07168ecfb4
|
|
@ -1,594 +0,0 @@
|
|||
import functools
|
||||
import os
|
||||
from collections import OrderedDict
|
||||
|
||||
import torch
|
||||
import torch.nn.functional as F
|
||||
import torchvision
|
||||
from torch import nn
|
||||
|
||||
from models.arch_util import ConvBnLelu, ConvGnSilu, ExpansionBlock, ExpansionBlock2, MultiConvBlock
|
||||
from models.switched_conv.switched_conv import BareConvSwitch, compute_attention_specificity, AttentionNorm
|
||||
from models.switched_conv.switched_conv_util import save_attention_to_image_rgb
|
||||
from trainer.networks import register_model
|
||||
from utils.util import checkpoint
|
||||
|
||||
|
||||
# VGG-style layer with Conv(stride2)->BN->Activation->Conv->BN->Activation
|
||||
# Doubles the input filter count.
|
||||
class HalvingProcessingBlock(nn.Module):
|
||||
def __init__(self, filters):
|
||||
super(HalvingProcessingBlock, self).__init__()
|
||||
self.bnconv1 = ConvGnSilu(filters, filters * 2, stride=2, norm=False, bias=False)
|
||||
self.bnconv2 = ConvGnSilu(filters * 2, filters * 2, norm=True, bias=False)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.bnconv1(x)
|
||||
return self.bnconv2(x)
|
||||
|
||||
|
||||
# This is a classic u-net architecture with the goal of assigning each individual pixel an individual transform
|
||||
# switching set.
|
||||
class ConvBasisMultiplexer(nn.Module):
|
||||
def __init__(self, input_channels, base_filters, reductions, processing_depth, multiplexer_channels, use_gn=True, use_exp2=False):
|
||||
super(ConvBasisMultiplexer, self).__init__()
|
||||
self.filter_conv = ConvGnSilu(input_channels, base_filters, bias=True)
|
||||
self.reduction_blocks = nn.ModuleList([HalvingProcessingBlock(base_filters * 2 ** i) for i in range(reductions)])
|
||||
reduction_filters = base_filters * 2 ** reductions
|
||||
self.processing_blocks = nn.Sequential(OrderedDict([('block%i' % (i,), ConvGnSilu(reduction_filters, reduction_filters, bias=False)) for i in range(processing_depth)]))
|
||||
if use_exp2:
|
||||
self.expansion_blocks = nn.ModuleList([ExpansionBlock2(reduction_filters // (2 ** i)) for i in range(reductions)])
|
||||
else:
|
||||
self.expansion_blocks = nn.ModuleList([ExpansionBlock(reduction_filters // (2 ** i)) for i in range(reductions)])
|
||||
|
||||
gap = base_filters - multiplexer_channels
|
||||
cbl1_out = ((base_filters - (gap // 2)) // 4) * 4 # Must be multiples of 4 to use with group norm.
|
||||
self.cbl1 = ConvGnSilu(base_filters, cbl1_out, norm=use_gn, bias=False, num_groups=4)
|
||||
cbl2_out = ((base_filters - (3 * gap // 4)) // 4) * 4
|
||||
self.cbl2 = ConvGnSilu(cbl1_out, cbl2_out, norm=use_gn, bias=False, num_groups=4)
|
||||
self.cbl3 = ConvGnSilu(cbl2_out, multiplexer_channels, bias=True, norm=False)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.filter_conv(x)
|
||||
reduction_identities = []
|
||||
for b in self.reduction_blocks:
|
||||
reduction_identities.append(x)
|
||||
x = b(x)
|
||||
x = self.processing_blocks(x)
|
||||
for i, b in enumerate(self.expansion_blocks):
|
||||
x = b(x, reduction_identities[-i - 1])
|
||||
|
||||
x = self.cbl1(x)
|
||||
x = self.cbl2(x)
|
||||
x = self.cbl3(x)
|
||||
return x
|
||||
|
||||
|
||||
# torch.gather() which operates across 2d images.
|
||||
def gather_2d(input, index):
|
||||
b, c, h, w = input.shape
|
||||
nodim = input.view(b, c, h * w)
|
||||
ind_nd = index[:, 0]*w + index[:, 1]
|
||||
ind_nd = ind_nd.unsqueeze(1)
|
||||
ind_nd = ind_nd.repeat((1, c))
|
||||
ind_nd = ind_nd.unsqueeze(2)
|
||||
result = torch.gather(nodim, dim=2, index=ind_nd)
|
||||
result = result.squeeze()
|
||||
if b == 1:
|
||||
result = result.unsqueeze(0)
|
||||
return result
|
||||
|
||||
|
||||
class ConfigurableSwitchComputer(nn.Module):
|
||||
def __init__(self, base_filters, multiplexer_net, pre_transform_block, transform_block, transform_count, attention_norm,
|
||||
post_transform_block=None,
|
||||
init_temp=20, add_scalable_noise_to_transforms=False, feed_transforms_into_multiplexer=False, post_switch_conv=True,
|
||||
anorm_multiplier=16):
|
||||
super(ConfigurableSwitchComputer, self).__init__()
|
||||
|
||||
tc = transform_count
|
||||
self.multiplexer = multiplexer_net(tc)
|
||||
|
||||
if pre_transform_block:
|
||||
self.pre_transform = pre_transform_block()
|
||||
else:
|
||||
self.pre_transform = None
|
||||
self.transforms = nn.ModuleList([transform_block() for _ in range(transform_count)])
|
||||
self.add_noise = add_scalable_noise_to_transforms
|
||||
self.feed_transforms_into_multiplexer = feed_transforms_into_multiplexer
|
||||
self.noise_scale = nn.Parameter(torch.full((1,), float(1e-3)))
|
||||
|
||||
# And the switch itself, including learned scalars
|
||||
self.switch = BareConvSwitch(initial_temperature=init_temp, attention_norm=AttentionNorm(transform_count, accumulator_size=anorm_multiplier * transform_count) if attention_norm else None)
|
||||
self.switch_scale = nn.Parameter(torch.full((1,), float(1)))
|
||||
self.post_transform_block = post_transform_block
|
||||
if post_switch_conv:
|
||||
self.post_switch_conv = ConvBnLelu(base_filters, base_filters, norm=False, bias=True)
|
||||
# The post_switch_conv gets a low scale initially. The network can decide to magnify it (or not)
|
||||
# depending on its needs.
|
||||
self.psc_scale = nn.Parameter(torch.full((1,), float(.1)))
|
||||
else:
|
||||
self.post_switch_conv = None
|
||||
self.update_norm = True
|
||||
|
||||
def set_update_attention_norm(self, set_val):
|
||||
self.update_norm = set_val
|
||||
|
||||
# Regarding inputs: it is acceptable to pass in a tuple/list as an input for (x), but the first element
|
||||
# *must* be the actual parameter that gets fed through the network - it is assumed to be the identity.
|
||||
def forward(self, x, att_in=None, identity=None, output_attention_weights=True, fixed_scale=1, do_checkpointing=False,
|
||||
output_att_logits=False):
|
||||
if isinstance(x, tuple):
|
||||
x1 = x[0]
|
||||
else:
|
||||
x1 = x
|
||||
|
||||
if att_in is None:
|
||||
att_in = x
|
||||
|
||||
if identity is None:
|
||||
identity = x1
|
||||
|
||||
if self.add_noise:
|
||||
rand_feature = torch.randn_like(x1) * self.noise_scale
|
||||
if isinstance(x, tuple):
|
||||
x = (x1 + rand_feature,) + x[1:]
|
||||
else:
|
||||
x = x1 + rand_feature
|
||||
|
||||
if not isinstance(x, tuple):
|
||||
x = (x,)
|
||||
if self.pre_transform:
|
||||
x = self.pre_transform(*x)
|
||||
if not isinstance(x, tuple):
|
||||
x = (x,)
|
||||
if do_checkpointing:
|
||||
xformed = [checkpoint(t, *x) for t in self.transforms]
|
||||
else:
|
||||
xformed = [t(*x) for t in self.transforms]
|
||||
|
||||
if not isinstance(att_in, tuple):
|
||||
att_in = (att_in,)
|
||||
if self.feed_transforms_into_multiplexer:
|
||||
att_in = att_in + (torch.stack(xformed, dim=1),)
|
||||
if do_checkpointing:
|
||||
m = checkpoint(self.multiplexer, *att_in)
|
||||
else:
|
||||
m = self.multiplexer(*att_in)
|
||||
|
||||
# It is assumed that [xformed] and [m] are collapsed into tensors at this point.
|
||||
outputs, attention, att_logits = self.switch(xformed, m, True, self.update_norm, output_attention_logits=True)
|
||||
if self.post_transform_block is not None:
|
||||
outputs = self.post_transform_block(outputs)
|
||||
|
||||
outputs = identity + outputs * self.switch_scale * fixed_scale
|
||||
if self.post_switch_conv is not None:
|
||||
outputs = outputs + self.post_switch_conv(outputs) * self.psc_scale * fixed_scale
|
||||
if output_attention_weights:
|
||||
if output_att_logits:
|
||||
return outputs, attention, att_logits
|
||||
else:
|
||||
return outputs, attention
|
||||
else:
|
||||
return outputs
|
||||
|
||||
def set_temperature(self, temp):
|
||||
self.switch.set_attention_temperature(temp)
|
||||
|
||||
|
||||
class ConfigurableSwitchedResidualGenerator2(nn.Module):
|
||||
def __init__(self, switch_depth, switch_filters, switch_reductions, switch_processing_layers, trans_counts, trans_kernel_sizes,
|
||||
trans_layers, transformation_filters, attention_norm, initial_temp=20, final_temperature_step=50000, heightened_temp_min=1,
|
||||
heightened_final_step=50000, upsample_factor=1,
|
||||
add_scalable_noise_to_transforms=False):
|
||||
super(ConfigurableSwitchedResidualGenerator2, self).__init__()
|
||||
switches = []
|
||||
self.initial_conv = ConvBnLelu(3, transformation_filters, norm=False, activation=False, bias=True)
|
||||
self.upconv1 = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.upconv2 = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.hr_conv = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.final_conv = ConvBnLelu(transformation_filters, 3, norm=False, activation=False, bias=True)
|
||||
for _ in range(switch_depth):
|
||||
multiplx_fn = functools.partial(ConvBasisMultiplexer, transformation_filters, switch_filters, switch_reductions, switch_processing_layers, trans_counts)
|
||||
pretransform_fn = functools.partial(ConvBnLelu, transformation_filters, transformation_filters, norm=False, bias=False, weight_init_factor=.1)
|
||||
transform_fn = functools.partial(MultiConvBlock, transformation_filters, int(transformation_filters * 1.5), transformation_filters, kernel_size=trans_kernel_sizes, depth=trans_layers, weight_init_factor=.1)
|
||||
switches.append(ConfigurableSwitchComputer(transformation_filters, multiplx_fn,
|
||||
pre_transform_block=pretransform_fn, transform_block=transform_fn,
|
||||
attention_norm=attention_norm,
|
||||
transform_count=trans_counts, init_temp=initial_temp,
|
||||
add_scalable_noise_to_transforms=add_scalable_noise_to_transforms))
|
||||
|
||||
self.switches = nn.ModuleList(switches)
|
||||
self.transformation_counts = trans_counts
|
||||
self.init_temperature = initial_temp
|
||||
self.final_temperature_step = final_temperature_step
|
||||
self.heightened_temp_min = heightened_temp_min
|
||||
self.heightened_final_step = heightened_final_step
|
||||
self.attentions = None
|
||||
self.upsample_factor = upsample_factor
|
||||
assert self.upsample_factor == 2 or self.upsample_factor == 4
|
||||
|
||||
def forward(self, x):
|
||||
# This is a common bug when evaluating SRG2 generators. It needs to be configured properly in eval mode. Just fail.
|
||||
if not self.train:
|
||||
assert self.switches[0].switch.temperature == 1
|
||||
|
||||
x = self.initial_conv(x)
|
||||
|
||||
self.attentions = []
|
||||
for i, sw in enumerate(self.switches):
|
||||
x, att = sw.forward(x, True)
|
||||
self.attentions.append(att)
|
||||
|
||||
x = self.upconv1(F.interpolate(x, scale_factor=2, mode="nearest"))
|
||||
if self.upsample_factor > 2:
|
||||
x = F.interpolate(x, scale_factor=2, mode="nearest")
|
||||
x = self.upconv2(x)
|
||||
x = self.final_conv(self.hr_conv(x))
|
||||
return x, x
|
||||
|
||||
def set_temperature(self, temp):
|
||||
[sw.set_temperature(temp) for sw in self.switches]
|
||||
|
||||
def update_for_step(self, step, experiments_path='.'):
|
||||
if self.attentions:
|
||||
temp = max(1,
|
||||
1 + self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step)
|
||||
if temp == 1 and self.heightened_final_step and step > self.final_temperature_step and \
|
||||
self.heightened_final_step != 1:
|
||||
# Once the temperature passes (1) it enters an inverted curve to match the linear curve from above.
|
||||
# without this, the attention specificity "spikes" incredibly fast in the last few iterations.
|
||||
h_steps_total = self.heightened_final_step - self.final_temperature_step
|
||||
h_steps_current = min(step - self.final_temperature_step, h_steps_total)
|
||||
# The "gap" will represent the steps that need to be traveled as a linear function.
|
||||
h_gap = 1 / self.heightened_temp_min
|
||||
temp = h_gap * h_steps_current / h_steps_total
|
||||
# Invert temperature to represent reality on this side of the curve
|
||||
temp = 1 / temp
|
||||
self.set_temperature(temp)
|
||||
if step % 50 == 0:
|
||||
output_path = os.path.join(experiments_path, "attention_maps", "a%i")
|
||||
prefix = "attention_map_%i_%%i.png" % (step,)
|
||||
[save_attention_to_image_rgb(output_path % (i,), self.attentions[i], self.transformation_counts, prefix, step) for i in range(len(self.attentions))]
|
||||
|
||||
def get_debug_values(self, step):
|
||||
temp = self.switches[0].switch.temperature
|
||||
mean_hists = [compute_attention_specificity(att, 2) for att in self.attentions]
|
||||
means = [i[0] for i in mean_hists]
|
||||
hists = [i[1].clone().detach().cpu().flatten() for i in mean_hists]
|
||||
val = {"switch_temperature": temp}
|
||||
for i in range(len(means)):
|
||||
val["switch_%i_specificity" % (i,)] = means[i]
|
||||
val["switch_%i_histogram" % (i,)] = hists[i]
|
||||
return val
|
||||
|
||||
|
||||
# Computes a linear latent by performing processing on the reference image and returning the filters of a single point,
|
||||
# which should be centered on the image patch being processed.
|
||||
#
|
||||
# Output is base_filters * 8.
|
||||
class ReferenceImageBranch(nn.Module):
|
||||
def __init__(self, base_filters=64):
|
||||
super(ReferenceImageBranch, self).__init__()
|
||||
self.features = nn.Sequential(ConvGnSilu(4, base_filters, kernel_size=7, bias=True),
|
||||
HalvingProcessingBlock(base_filters),
|
||||
ConvGnSilu(base_filters*2, base_filters*2, activation=True, norm=True, bias=False),
|
||||
HalvingProcessingBlock(base_filters*2),
|
||||
ConvGnSilu(base_filters*4, base_filters*4, activation=True, norm=True, bias=False),
|
||||
HalvingProcessingBlock(base_filters*4),
|
||||
ConvGnSilu(base_filters*8, base_filters*8, activation=True, norm=True, bias=False),
|
||||
ConvGnSilu(base_filters*8, base_filters*8, activation=True, norm=True, bias=False))
|
||||
|
||||
# center_point is a [b,2] long tensor describing the center point of where the patch was taken from the reference
|
||||
# image.
|
||||
def forward(self, x, center_point):
|
||||
x = self.features(x)
|
||||
return gather_2d(x, center_point // 8) # Divide by 8 to scale the center_point down.
|
||||
|
||||
|
||||
# Mutiplexer that combines a structured embedding with a contextual switch input to guide alterations to that input.
|
||||
#
|
||||
# Implemented as basically a u-net which reduces the input into the same structural space as the embedding, combines the
|
||||
# two, then expands back into the original feature space.
|
||||
class EmbeddingMultiplexer(nn.Module):
|
||||
# Note: reductions=2 if the encoder is using interpolated input, otherwise reductions=3.
|
||||
def __init__(self, nf, multiplexer_channels, reductions=2):
|
||||
super(EmbeddingMultiplexer, self).__init__()
|
||||
self.embedding_process = MultiConvBlock(256, 256, 256, kernel_size=3, depth=3, norm=True)
|
||||
|
||||
self.filter_conv = ConvGnSilu(nf, nf, activation=True, norm=False, bias=True)
|
||||
self.reduction_blocks = nn.ModuleList([HalvingProcessingBlock(nf * 2 ** i) for i in range(reductions)])
|
||||
reduction_filters = nf * 2 ** reductions
|
||||
self.processing_blocks = nn.Sequential(
|
||||
ConvGnSilu(reduction_filters + 256, reduction_filters + 256, kernel_size=1, activation=True, norm=False, bias=True),
|
||||
ConvGnSilu(reduction_filters + 256, reduction_filters + 128, kernel_size=1, activation=True, norm=True, bias=False),
|
||||
ConvGnSilu(reduction_filters + 128, reduction_filters, kernel_size=3, activation=True, norm=True, bias=False),
|
||||
ConvGnSilu(reduction_filters, reduction_filters, kernel_size=3, activation=True, norm=True, bias=False))
|
||||
self.expansion_blocks = nn.ModuleList([ExpansionBlock2(reduction_filters // (2 ** i)) for i in range(reductions)])
|
||||
|
||||
gap = nf - multiplexer_channels
|
||||
cbl1_out = ((nf - (gap // 2)) // 4) * 4 # Must be multiples of 4 to use with group norm.
|
||||
self.cbl1 = ConvGnSilu(nf, cbl1_out, norm=True, bias=False, num_groups=4)
|
||||
cbl2_out = ((nf - (3 * gap // 4)) // 4) * 4
|
||||
self.cbl2 = ConvGnSilu(cbl1_out, cbl2_out, norm=True, bias=False, num_groups=4)
|
||||
self.cbl3 = ConvGnSilu(cbl2_out, multiplexer_channels, bias=True, norm=False)
|
||||
|
||||
def forward(self, x, embedding):
|
||||
x = self.filter_conv(x)
|
||||
embedding = self.embedding_process(embedding)
|
||||
|
||||
reduction_identities = []
|
||||
for b in self.reduction_blocks:
|
||||
reduction_identities.append(x)
|
||||
x = b(x)
|
||||
x = self.processing_blocks(torch.cat([x, embedding], dim=1))
|
||||
for i, b in enumerate(self.expansion_blocks):
|
||||
x = b(x, reduction_identities[-i - 1])
|
||||
|
||||
x = self.cbl1(x)
|
||||
x = self.cbl2(x)
|
||||
x = self.cbl3(x)
|
||||
return x
|
||||
|
||||
|
||||
class QueryKeyMultiplexer(nn.Module):
|
||||
def __init__(self, nf, multiplexer_channels, embedding_channels=256, reductions=2):
|
||||
super(QueryKeyMultiplexer, self).__init__()
|
||||
|
||||
# Blocks used to create the query
|
||||
self.input_process = ConvGnSilu(nf, nf, activation=True, norm=False, bias=True)
|
||||
self.embedding_process = ConvGnSilu(embedding_channels, 256, activation=True, norm=False, bias=True)
|
||||
self.reduction_blocks = nn.ModuleList([HalvingProcessingBlock(nf * 2 ** i) for i in range(reductions)])
|
||||
reduction_filters = nf * 2 ** reductions
|
||||
self.processing_blocks = nn.Sequential(
|
||||
ConvGnSilu(reduction_filters + 256, reduction_filters + 256, kernel_size=1, activation=True, norm=False, bias=True),
|
||||
ConvGnSilu(reduction_filters + 256, reduction_filters + 128, kernel_size=1, activation=True, norm=True, bias=False),
|
||||
ConvGnSilu(reduction_filters + 128, reduction_filters, kernel_size=3, activation=True, norm=True, bias=False),
|
||||
ConvGnSilu(reduction_filters, reduction_filters, kernel_size=3, activation=True, norm=True, bias=False))
|
||||
self.expansion_blocks = nn.ModuleList([ExpansionBlock2(reduction_filters // (2 ** i)) for i in range(reductions)])
|
||||
|
||||
# Blocks used to create the key
|
||||
self.key_process = ConvGnSilu(nf, nf, kernel_size=1, activation=True, norm=False, bias=True)
|
||||
|
||||
# Postprocessing blocks.
|
||||
self.query_key_combine = ConvGnSilu(nf*2, nf, kernel_size=1, activation=True, norm=False, bias=False)
|
||||
self.cbl1 = ConvGnSilu(nf, nf // 2, kernel_size=1, norm=True, bias=False, num_groups=4)
|
||||
self.cbl2 = ConvGnSilu(nf // 2, 1, kernel_size=1, norm=False, bias=False)
|
||||
|
||||
def forward(self, x, embedding, transformations):
|
||||
q = self.input_process(x)
|
||||
embedding = self.embedding_process(embedding)
|
||||
reduction_identities = []
|
||||
for b in self.reduction_blocks:
|
||||
reduction_identities.append(q)
|
||||
q = b(q)
|
||||
q = self.processing_blocks(torch.cat([q, embedding], dim=1))
|
||||
for i, b in enumerate(self.expansion_blocks):
|
||||
q = b(q, reduction_identities[-i - 1])
|
||||
|
||||
b, t, f, h, w = transformations.shape
|
||||
k = transformations.view(b * t, f, h, w)
|
||||
k = self.key_process(k)
|
||||
|
||||
q = q.view(b, 1, f, h, w).repeat(1, t, 1, 1, 1).view(b * t, f, h, w)
|
||||
v = self.query_key_combine(torch.cat([q, k], dim=1))
|
||||
|
||||
v = self.cbl1(v)
|
||||
v = self.cbl2(v)
|
||||
|
||||
return v.view(b, t, h, w)
|
||||
|
||||
|
||||
class QueryKeyPyramidMultiplexer(nn.Module):
|
||||
def __init__(self, nf, multiplexer_channels, reductions=3):
|
||||
super(QueryKeyPyramidMultiplexer, self).__init__()
|
||||
|
||||
# Blocks used to create the query
|
||||
self.input_process = ConvGnSilu(nf, nf, activation=True, norm=False, bias=True)
|
||||
self.reduction_blocks = nn.ModuleList([HalvingProcessingBlock(nf * 2 ** i) for i in range(reductions)])
|
||||
reduction_filters = nf * 2 ** reductions
|
||||
self.processing_blocks = nn.Sequential(OrderedDict([('block%i' % (i,), ConvGnSilu(reduction_filters, reduction_filters, kernel_size=1, norm=True, bias=False)) for i in range(3)]))
|
||||
self.expansion_blocks = nn.ModuleList([ExpansionBlock2(reduction_filters // (2 ** i)) for i in range(reductions)])
|
||||
|
||||
# Blocks used to create the key
|
||||
self.key_process = ConvGnSilu(nf, nf, kernel_size=1, activation=True, norm=False, bias=True)
|
||||
|
||||
# Postprocessing blocks.
|
||||
self.query_key_combine = ConvGnSilu(nf*2, nf, kernel_size=3, activation=True, norm=False, bias=False)
|
||||
self.cbl0 = ConvGnSilu(nf, nf, kernel_size=3, activation=True, norm=True, bias=False)
|
||||
self.cbl1 = ConvGnSilu(nf, nf // 2, kernel_size=1, norm=True, bias=False, num_groups=4)
|
||||
self.cbl2 = ConvGnSilu(nf // 2, 1, kernel_size=1, norm=False, bias=False)
|
||||
|
||||
def forward(self, x, transformations):
|
||||
q = self.input_process(x)
|
||||
reduction_identities = []
|
||||
for b in self.reduction_blocks:
|
||||
reduction_identities.append(q)
|
||||
q = b(q)
|
||||
q = self.processing_blocks(q)
|
||||
for i, b in enumerate(self.expansion_blocks):
|
||||
q = b(q, reduction_identities[-i - 1])
|
||||
|
||||
b, t, f, h, w = transformations.shape
|
||||
k = transformations.view(b * t, f, h, w)
|
||||
k = self.key_process(k)
|
||||
|
||||
q = q.view(b, 1, f, h, w).repeat(1, t, 1, 1, 1).view(b * t, f, h, w)
|
||||
v = self.query_key_combine(torch.cat([q, k], dim=1))
|
||||
v = self.cbl0(v)
|
||||
v = self.cbl1(v)
|
||||
v = self.cbl2(v)
|
||||
|
||||
return v.view(b, t, h, w)
|
||||
|
||||
|
||||
# Base class for models that utilize ConfigurableSwitchComputer. Provides basis functionality like logging
|
||||
# switch temperature, distribution and images, as well as managing attention norms.
|
||||
class SwitchModelBase(nn.Module):
|
||||
def __init__(self, init_temperature=10, final_temperature_step=10000):
|
||||
super(SwitchModelBase, self).__init__()
|
||||
self.switches = [] # The implementing class is expected to set this to a list of all ConfigurableSwitchComputers.
|
||||
self.attentions = [] # The implementing class is expected to set this in forward() to the output of the attention blocks.
|
||||
self.lr = None # The implementing class is expected to set this to the input image fed into the generator. If not
|
||||
# set, the attention logger will not output an image reference.
|
||||
self.init_temperature = init_temperature
|
||||
self.final_temperature_step = final_temperature_step
|
||||
|
||||
def set_temperature(self, temp):
|
||||
[sw.set_temperature(temp) for sw in self.switches]
|
||||
|
||||
def update_for_step(self, step, experiments_path='.'):
|
||||
# All-reduce the attention norm.
|
||||
for sw in self.switches:
|
||||
sw.switch.reduce_norm_params()
|
||||
|
||||
temp = max(1, 1 + self.init_temperature *
|
||||
(self.final_temperature_step - step) / self.final_temperature_step)
|
||||
self.set_temperature(temp)
|
||||
if step % 100 == 0:
|
||||
output_path = os.path.join(experiments_path, "attention_maps")
|
||||
prefix = "amap_%i_a%i_%%i.png"
|
||||
[save_attention_to_image_rgb(output_path, self.attentions[i], self.attentions[i].shape[3], prefix % (step, i), step,
|
||||
output_mag=False) for i in range(len(self.attentions))]
|
||||
if self.lr is not None:
|
||||
torchvision.utils.save_image(self.lr[:, :3], os.path.join(experiments_path, "attention_maps",
|
||||
"amap_%i_base_image.png" % (step,)))
|
||||
|
||||
# This is a bit awkward. We want this plot to show up in TB as a histogram, but we are getting an intensity
|
||||
# plot out of the attention norm tensor. So we need to convert it back into a list of indexes, then feed into TB.
|
||||
def compute_anorm_histogram(self):
|
||||
intensities = [sw.switch.attention_norm.compute_buffer_norm().clone().detach().cpu() for sw in self.switches]
|
||||
result = []
|
||||
for intensity in intensities:
|
||||
intensity = intensity * 10
|
||||
bins = torch.tensor(list(range(len(intensity))))
|
||||
intensity = intensity.long()
|
||||
result.append(bins.repeat_interleave(intensity, 0))
|
||||
return result
|
||||
|
||||
def get_debug_values(self, step, net_name):
|
||||
temp = self.switches[0].switch.temperature
|
||||
mean_hists = [compute_attention_specificity(att, 2) for att in self.attentions]
|
||||
means = [i[0] for i in mean_hists]
|
||||
hists = [i[1].clone().detach().cpu().flatten() for i in mean_hists]
|
||||
anorms = self.compute_anorm_histogram()
|
||||
val = {"switch_temperature": temp}
|
||||
for i in range(len(means)):
|
||||
val["switch_%i_specificity" % (i,)] = means[i]
|
||||
val["switch_%i_histogram" % (i,)] = hists[i]
|
||||
val["switch_%i_attention_norm_histogram" % (i,)] = anorms[i]
|
||||
return val
|
||||
|
||||
|
||||
class ConfigurableSwitchedResidualGenerator2(nn.Module):
|
||||
def __init__(self, switch_depth, switch_filters, switch_reductions, switch_processing_layers, trans_counts, trans_kernel_sizes,
|
||||
trans_layers, transformation_filters, attention_norm, initial_temp=20, final_temperature_step=50000, heightened_temp_min=1,
|
||||
heightened_final_step=50000, upsample_factor=1,
|
||||
add_scalable_noise_to_transforms=False):
|
||||
super(ConfigurableSwitchedResidualGenerator2, self).__init__()
|
||||
switches = []
|
||||
self.initial_conv = ConvBnLelu(3, transformation_filters, kernel_size=7, norm=False, activation=False, bias=True)
|
||||
self.upconv1 = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.upconv2 = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.hr_conv = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.final_conv = ConvBnLelu(transformation_filters, 3, norm=False, activation=False, bias=True)
|
||||
for _ in range(switch_depth):
|
||||
multiplx_fn = functools.partial(ConvBasisMultiplexer, transformation_filters, switch_filters, switch_reductions, switch_processing_layers, trans_counts)
|
||||
pretransform_fn = functools.partial(ConvBnLelu, transformation_filters, transformation_filters, norm=False, bias=False, weight_init_factor=.1)
|
||||
transform_fn = functools.partial(MultiConvBlock, transformation_filters, int(transformation_filters * 1.5), transformation_filters, kernel_size=trans_kernel_sizes, depth=trans_layers, weight_init_factor=.1)
|
||||
switches.append(ConfigurableSwitchComputer(transformation_filters, multiplx_fn,
|
||||
pre_transform_block=pretransform_fn, transform_block=transform_fn,
|
||||
attention_norm=attention_norm,
|
||||
transform_count=trans_counts, init_temp=initial_temp,
|
||||
add_scalable_noise_to_transforms=add_scalable_noise_to_transforms))
|
||||
|
||||
self.switches = nn.ModuleList(switches)
|
||||
self.transformation_counts = trans_counts
|
||||
self.init_temperature = initial_temp
|
||||
self.final_temperature_step = final_temperature_step
|
||||
self.heightened_temp_min = heightened_temp_min
|
||||
self.heightened_final_step = heightened_final_step
|
||||
self.attentions = None
|
||||
self.upsample_factor = upsample_factor
|
||||
self.lr = None
|
||||
assert self.upsample_factor == 2 or self.upsample_factor == 4
|
||||
|
||||
def forward(self, x):
|
||||
self.lr = x.detach().cpu()
|
||||
|
||||
# This is a common bug when evaluating SRG2 generators. It needs to be configured properly in eval mode. Just fail.
|
||||
if not self.train:
|
||||
assert self.switches[0].switch.temperature == 1
|
||||
|
||||
x = self.initial_conv(x)
|
||||
|
||||
self.attentions = []
|
||||
for i, sw in enumerate(self.switches):
|
||||
x, att = checkpoint(sw, x)
|
||||
self.attentions.append(att)
|
||||
|
||||
x = self.upconv1(F.interpolate(x, scale_factor=2, mode="nearest"))
|
||||
if self.upsample_factor > 2:
|
||||
x = F.interpolate(x, scale_factor=2, mode="nearest")
|
||||
x = self.upconv2(x)
|
||||
x = self.final_conv(self.hr_conv(x))
|
||||
return x
|
||||
|
||||
def set_temperature(self, temp):
|
||||
[sw.set_temperature(temp) for sw in self.switches]
|
||||
|
||||
def update_for_step(self, step, experiments_path='.'):
|
||||
if self.attentions:
|
||||
temp = max(1,
|
||||
1 + self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step)
|
||||
if temp == 1 and self.heightened_final_step and step > self.final_temperature_step and \
|
||||
self.heightened_final_step != 1:
|
||||
# Once the temperature passes (1) it enters an inverted curve to match the linear curve from above.
|
||||
# without this, the attention specificity "spikes" incredibly fast in the last few iterations.
|
||||
h_steps_total = self.heightened_final_step - self.final_temperature_step
|
||||
h_steps_current = min(step - self.final_temperature_step, h_steps_total)
|
||||
# The "gap" will represent the steps that need to be traveled as a linear function.
|
||||
h_gap = 1 / self.heightened_temp_min
|
||||
temp = h_gap * h_steps_current / h_steps_total
|
||||
# Invert temperature to represent reality on this side of the curve
|
||||
temp = 1 / temp
|
||||
self.set_temperature(temp)
|
||||
if step % 100 == 0:
|
||||
output_path = os.path.join(experiments_path, "attention_maps")
|
||||
prefix = "amap_%i_a%i_%%i.png"
|
||||
[save_attention_to_image_rgb(output_path, self.attentions[i], self.attentions[i].shape[3], prefix % (step, i), step,
|
||||
output_mag=False) for i in range(len(self.attentions))]
|
||||
if self.lr is not None:
|
||||
torchvision.utils.save_image(self.lr[:, :3], os.path.join(experiments_path, "attention_maps",
|
||||
"amap_%i_base_image.png" % (step,)))
|
||||
|
||||
def get_debug_values(self, step, net_name):
|
||||
temp = self.switches[0].switch.temperature
|
||||
mean_hists = [compute_attention_specificity(att, 2) for att in self.attentions]
|
||||
means = [i[0] for i in mean_hists]
|
||||
hists = [i[1].clone().detach().cpu().flatten() for i in mean_hists]
|
||||
val = {"switch_temperature": temp}
|
||||
for i in range(len(means)):
|
||||
val["switch_%i_specificity" % (i,)] = means[i]
|
||||
val["switch_%i_histogram" % (i,)] = hists[i]
|
||||
return val
|
||||
|
||||
@register_model
|
||||
def register_ConfigurableSwitchedResidualGenerator2(opt_net, opt):
|
||||
return ConfigurableSwitchedResidualGenerator2(switch_depth=opt_net['switch_depth'],
|
||||
switch_filters=opt_net['switch_filters'],
|
||||
switch_reductions=opt_net['switch_reductions'],
|
||||
switch_processing_layers=opt_net[
|
||||
'switch_processing_layers'],
|
||||
trans_counts=opt_net['trans_counts'],
|
||||
trans_kernel_sizes=opt_net['trans_kernel_sizes'],
|
||||
trans_layers=opt_net['trans_layers'],
|
||||
transformation_filters=opt_net['transformation_filters'],
|
||||
attention_norm=opt_net['attention_norm'],
|
||||
initial_temp=opt_net['temperature'],
|
||||
final_temperature_step=opt_net['temperature_final_step'],
|
||||
heightened_temp_min=opt_net['heightened_temp_min'],
|
||||
heightened_final_step=opt_net['heightened_final_step'],
|
||||
upsample_factor=scale,
|
||||
add_scalable_noise_to_transforms=opt_net['add_noise'],
|
||||
for_video=opt_net['for_video'])
|
||||
|
|
@ -1,223 +0,0 @@
|
|||
import os
|
||||
|
||||
import torch
|
||||
import torchvision
|
||||
from matplotlib import cm
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
import functools
|
||||
from collections import OrderedDict
|
||||
|
||||
from models.SwitchedResidualGenerator_arch import HalvingProcessingBlock, ConfigurableSwitchComputer
|
||||
from models.arch_util import ConvBnLelu, ConvGnSilu, ExpansionBlock, MultiConvBlock
|
||||
from models.switched_conv.switched_conv import BareConvSwitch, AttentionNorm
|
||||
from trainer.networks import register_model
|
||||
from utils.util import checkpoint
|
||||
|
||||
|
||||
# This is a classic u-net architecture with the goal of assigning each individual pixel an individual transform
|
||||
# switching set.
|
||||
class ConvBasisMultiplexer(nn.Module):
|
||||
def __init__(self, input_channels, base_filters, reductions, processing_depth, multiplexer_channels, use_gn=True):
|
||||
super(ConvBasisMultiplexer, self).__init__()
|
||||
self.filter_conv = ConvGnSilu(input_channels, base_filters, bias=True)
|
||||
self.reduction_blocks = nn.ModuleList([HalvingProcessingBlock(base_filters * 2 ** i) for i in range(reductions)])
|
||||
reduction_filters = base_filters * 2 ** reductions
|
||||
self.processing_blocks = nn.Sequential(OrderedDict([('block%i' % (i,), ConvGnSilu(reduction_filters, reduction_filters, bias=False)) for i in range(processing_depth)]))
|
||||
self.expansion_blocks = nn.ModuleList([ExpansionBlock(reduction_filters // (2 ** i)) for i in range(reductions)])
|
||||
|
||||
gap = base_filters - multiplexer_channels
|
||||
cbl1_out = ((base_filters - (gap // 2)) // 4) * 4 # Must be multiples of 4 to use with group norm.
|
||||
self.cbl1 = ConvGnSilu(base_filters, cbl1_out, norm=use_gn, bias=False, num_groups=4)
|
||||
cbl2_out = ((base_filters - (3 * gap // 4)) // 4) * 4
|
||||
self.cbl2 = ConvGnSilu(cbl1_out, cbl2_out, norm=use_gn, bias=False, num_groups=4)
|
||||
self.cbl3 = ConvGnSilu(cbl2_out, multiplexer_channels, bias=True, norm=False)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.filter_conv(x)
|
||||
reduction_identities = []
|
||||
for b in self.reduction_blocks:
|
||||
reduction_identities.append(x)
|
||||
x = b(x)
|
||||
x = self.processing_blocks(x)
|
||||
for i, b in enumerate(self.expansion_blocks):
|
||||
x = b(x, reduction_identities[-i - 1])
|
||||
|
||||
x = self.cbl1(x)
|
||||
x = self.cbl2(x)
|
||||
x = self.cbl3(x)
|
||||
return x
|
||||
|
||||
|
||||
def compute_attention_specificity(att_weights, topk=3):
|
||||
att = att_weights.detach()
|
||||
vals, indices = torch.topk(att, topk, dim=-1)
|
||||
avg = torch.sum(vals, dim=-1)
|
||||
avg = avg.flatten().mean()
|
||||
return avg.item(), indices.flatten().detach()
|
||||
|
||||
|
||||
# Copied from torchvision.utils.save_image. Allows specifying pixel format.
|
||||
def save_image(tensor, fp, nrow=8, padding=2,
|
||||
normalize=False, range=None, scale_each=False, pad_value=0, format=None, pix_format=None):
|
||||
from PIL import Image
|
||||
grid = torchvision.utils.make_grid(tensor, nrow=nrow, padding=padding, pad_value=pad_value,
|
||||
normalize=normalize, range=range, scale_each=scale_each)
|
||||
# Add 0.5 after unnormalizing to [0, 255] to round to nearest integer
|
||||
ndarr = grid.mul(255).add_(0.5).clamp_(0, 255).permute(1, 2, 0).to('cpu', torch.uint8).numpy()
|
||||
im = Image.fromarray(ndarr, mode=pix_format).convert('RGB')
|
||||
im.save(fp, format=format)
|
||||
|
||||
|
||||
def save_attention_to_image(folder, attention_out, attention_size, step, fname_part="map", l_mult=1.0):
|
||||
magnitude, indices = torch.topk(attention_out, 1, dim=-1)
|
||||
magnitude = magnitude.squeeze(3)
|
||||
indices = indices.squeeze(3)
|
||||
# indices is an integer tensor (b,w,h) where values are on the range [0,attention_size]
|
||||
# magnitude is a float tensor (b,w,h) [0,1] representing the magnitude of that attention.
|
||||
# Use HSV colorspace to show this. Hue is mapped to the indices, Lightness is mapped to intensity,
|
||||
# Saturation is left fixed.
|
||||
hue = indices.float() / attention_size
|
||||
saturation = torch.full_like(hue, .8)
|
||||
value = magnitude * l_mult
|
||||
hsv_img = torch.stack([hue, saturation, value], dim=1)
|
||||
|
||||
output_path=os.path.join(folder, "attention_maps", fname_part)
|
||||
os.makedirs(output_path, exist_ok=True)
|
||||
save_image(hsv_img, os.path.join(output_path, "attention_map_%i.png" % (step,)), pix_format="HSV")
|
||||
|
||||
|
||||
def save_attention_to_image_rgb(output_folder, attention_out, attention_size, file_prefix, step, cmap_discrete_name='viridis'):
|
||||
magnitude, indices = torch.topk(attention_out, 3, dim=-1)
|
||||
magnitude = magnitude.cpu()
|
||||
indices = indices.cpu()
|
||||
magnitude /= torch.max(torch.abs(torch.min(magnitude)), torch.abs(torch.max(magnitude)))
|
||||
colormap = cm.get_cmap(cmap_discrete_name, attention_size)
|
||||
colormap_mag = cm.get_cmap(cmap_discrete_name)
|
||||
os.makedirs(os.path.join(output_folder), exist_ok=True)
|
||||
for i in range(3):
|
||||
img = torch.tensor(colormap(indices[:,:,:,i].detach().numpy()))
|
||||
img = img.permute((0, 3, 1, 2))
|
||||
save_image(img, os.path.join(output_folder, file_prefix + "_%i_%s.png" % (step, "rgb_%i" % (i,))), pix_format="RGBA")
|
||||
|
||||
mag_image = torch.tensor(colormap_mag(magnitude[:,:,:,i].detach().numpy()))
|
||||
mag_image = mag_image.permute((0, 3, 1, 2))
|
||||
save_image(mag_image, os.path.join(output_folder, file_prefix + "_%i_%s.png" % (step, "mag_%i" % (i,))), pix_format="RGBA")
|
||||
|
||||
|
||||
class ConfigurableSwitchedResidualGenerator2(nn.Module):
|
||||
def __init__(self, switch_depth, switch_filters, switch_reductions, switch_processing_layers, trans_counts, trans_kernel_sizes,
|
||||
trans_layers, transformation_filters, initial_temp=20, final_temperature_step=50000, heightened_temp_min=1,
|
||||
heightened_final_step=50000, upsample_factor=1,
|
||||
add_scalable_noise_to_transforms=False, for_video=False):
|
||||
super(ConfigurableSwitchedResidualGenerator2, self).__init__()
|
||||
switches = []
|
||||
self.for_video = for_video
|
||||
if for_video:
|
||||
self.initial_conv = ConvBnLelu(6, transformation_filters, stride=upsample_factor, norm=False, activation=False, bias=True)
|
||||
else:
|
||||
self.initial_conv = ConvBnLelu(3, transformation_filters, norm=False, activation=False, bias=True)
|
||||
self.upconv1 = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.upconv2 = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.hr_conv = ConvBnLelu(transformation_filters, transformation_filters, norm=False, bias=True)
|
||||
self.final_conv = ConvBnLelu(transformation_filters, 3, norm=False, activation=False, bias=True)
|
||||
for _ in range(switch_depth):
|
||||
multiplx_fn = functools.partial(ConvBasisMultiplexer, transformation_filters, switch_filters, switch_reductions, switch_processing_layers, trans_counts)
|
||||
pretransform_fn = functools.partial(ConvBnLelu, transformation_filters, transformation_filters, norm=False, bias=False, weight_init_factor=.1)
|
||||
transform_fn = functools.partial(MultiConvBlock, transformation_filters, int(transformation_filters * 1.5), transformation_filters, kernel_size=trans_kernel_sizes, depth=trans_layers, weight_init_factor=.1)
|
||||
switches.append(ConfigurableSwitchComputer(transformation_filters, multiplx_fn, attention_norm=True,
|
||||
pre_transform_block=pretransform_fn, transform_block=transform_fn,
|
||||
transform_count=trans_counts, init_temp=initial_temp,
|
||||
add_scalable_noise_to_transforms=add_scalable_noise_to_transforms))
|
||||
|
||||
self.switches = nn.ModuleList(switches)
|
||||
self.transformation_counts = trans_counts
|
||||
self.init_temperature = initial_temp
|
||||
self.final_temperature_step = final_temperature_step
|
||||
self.heightened_temp_min = heightened_temp_min
|
||||
self.heightened_final_step = heightened_final_step
|
||||
self.attentions = None
|
||||
self.upsample_factor = upsample_factor
|
||||
assert self.upsample_factor == 2 or self.upsample_factor == 4
|
||||
|
||||
def forward(self, x, ref=None):
|
||||
if self.for_video:
|
||||
x_lg = F.interpolate(x, scale_factor=self.upsample_factor, mode="bicubic")
|
||||
if ref is None:
|
||||
ref = torch.zeros_like(x_lg)
|
||||
x_lg = torch.cat([x_lg, ref], dim=1)
|
||||
else:
|
||||
x_lg = x
|
||||
x = self.initial_conv(x_lg)
|
||||
|
||||
self.attentions = []
|
||||
for i, sw in enumerate(self.switches):
|
||||
x, att = checkpoint(sw, x)
|
||||
self.attentions.append(att)
|
||||
|
||||
x = self.upconv1(F.interpolate(x, scale_factor=2, mode="nearest"))
|
||||
if self.upsample_factor > 2:
|
||||
x = F.interpolate(x, scale_factor=2, mode="nearest")
|
||||
x = self.upconv2(x)
|
||||
x = self.final_conv(self.hr_conv(x))
|
||||
return x
|
||||
|
||||
def set_temperature(self, temp):
|
||||
[sw.set_temperature(temp) for sw in self.switches]
|
||||
|
||||
def update_for_step(self, step, experiments_path='.'):
|
||||
if self.attentions:
|
||||
temp = max(1,
|
||||
1 + self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step)
|
||||
if temp == 1 and self.heightened_final_step and step > self.final_temperature_step and \
|
||||
self.heightened_final_step != 1:
|
||||
# Once the temperature passes (1) it enters an inverted curve to match the linear curve from above.
|
||||
# without this, the attention specificity "spikes" incredibly fast in the last few iterations.
|
||||
h_steps_total = self.heightened_final_step - self.final_temperature_step
|
||||
h_steps_current = min(step - self.final_temperature_step, h_steps_total)
|
||||
# The "gap" will represent the steps that need to be traveled as a linear function.
|
||||
h_gap = 1 / self.heightened_temp_min
|
||||
temp = h_gap * h_steps_current / h_steps_total
|
||||
# Invert temperature to represent reality on this side of the curve
|
||||
temp = 1 / temp
|
||||
self.set_temperature(temp)
|
||||
if step % 50 == 0:
|
||||
[save_attention_to_image(experiments_path, self.attentions[i], self.transformation_counts, step, "a%i" % (i+1,), l_mult=10) for i in range(len(self.attentions))]
|
||||
|
||||
def get_debug_values(self, step, net_name):
|
||||
temp = self.switches[0].switch.temperature
|
||||
mean_hists = [compute_attention_specificity(att, 2) for att in self.attentions]
|
||||
means = [i[0] for i in mean_hists]
|
||||
hists = [i[1].clone().detach().cpu().flatten() for i in mean_hists]
|
||||
val = {"switch_temperature": temp}
|
||||
for i in range(len(means)):
|
||||
val["switch_%i_specificity" % (i,)] = means[i]
|
||||
val["switch_%i_histogram" % (i,)] = hists[i]
|
||||
return val
|
||||
|
||||
|
||||
class Interpolate(nn.Module):
|
||||
def __init__(self, factor, mode="nearest"):
|
||||
super(Interpolate, self).__init__()
|
||||
self.factor = factor
|
||||
self.mode = mode
|
||||
|
||||
def forward(self, x):
|
||||
return F.interpolate(x, scale_factor=self.factor, mode=self.mode)
|
||||
|
||||
@register_model
|
||||
def register_srg2classic(opt_net, opt):
|
||||
return ConfigurableSwitchedResidualGenerator2(switch_depth=opt_net['switch_depth'],
|
||||
switch_filters=opt_net['switch_filters'],
|
||||
switch_reductions=opt_net['switch_reductions'],
|
||||
switch_processing_layers=opt_net['switch_processing_layers'],
|
||||
trans_counts=opt_net['trans_counts'],
|
||||
trans_kernel_sizes=opt_net['trans_kernel_sizes'],
|
||||
trans_layers=opt_net['trans_layers'],
|
||||
transformation_filters=opt_net['transformation_filters'],
|
||||
initial_temp=opt_net['temperature'],
|
||||
final_temperature_step=opt_net['temperature_final_step'],
|
||||
heightened_temp_min=opt_net['heightened_temp_min'],
|
||||
heightened_final_step=opt_net['heightened_final_step'],
|
||||
upsample_factor=scale,
|
||||
add_scalable_noise_to_transforms=opt_net['add_noise'])
|
||||
|
|
@ -1 +0,0 @@
|
|||
Subproject commit cb520afd4da97796bfca398feeef18a7bd18475c
|
||||
48
codes/models/vqvae/kmeans_mask_producer.py
Normal file
48
codes/models/vqvae/kmeans_mask_producer.py
Normal file
|
|
@ -0,0 +1,48 @@
|
|||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
from torchvision.models.resnet import Bottleneck
|
||||
|
||||
from models.pixel_level_contrastive_learning.resnet_unet import UResNet50
|
||||
from trainer.networks import register_model
|
||||
from utils.kmeans import kmeans_predict
|
||||
from utils.util import opt_get
|
||||
|
||||
|
||||
class UResnetMaskProducer(nn.Module):
|
||||
def __init__(self, pretrained_uresnet_path, kmeans_centroid_path, mask_scales=[.125,.25,.5,1]):
|
||||
super().__init__()
|
||||
_, centroids = torch.load(kmeans_centroid_path)
|
||||
self.centroids = nn.Parameter(centroids)
|
||||
self.ures = UResNet50(Bottleneck, [3,4,6,3], out_dim=512).to('cuda')
|
||||
self.mask_scales = mask_scales
|
||||
|
||||
sd = torch.load(pretrained_uresnet_path)
|
||||
# An assumption is made that the state_dict came from a byol model. Strip out unnecessary weights..
|
||||
resnet_sd = {}
|
||||
for k, v in sd.items():
|
||||
if 'target_encoder.net.' in k:
|
||||
resnet_sd[k.replace('target_encoder.net.', '')] = v
|
||||
|
||||
self.ures.load_state_dict(resnet_sd, strict=True)
|
||||
self.ures.eval()
|
||||
|
||||
def forward(self, x):
|
||||
with torch.no_grad():
|
||||
latents = self.ures(x)
|
||||
b,c,h,w = latents.shape
|
||||
latents = latents.permute(0,2,3,1).reshape(b*h*w,c)
|
||||
masks = kmeans_predict(latents, self.centroids).float()
|
||||
masks = masks.reshape(b,1,h,w)
|
||||
interpolated_masks = {}
|
||||
for sf in self.mask_scales:
|
||||
dim_h, dim_w = int(sf*x.shape[-2]), int(sf*x.shape[-1])
|
||||
imask = F.interpolate(masks, size=(dim_h,dim_w), mode="nearest")
|
||||
interpolated_masks[dim_w] = imask.long()
|
||||
return interpolated_masks
|
||||
|
||||
|
||||
@register_model
|
||||
def register_uresnet_mask_producer(opt_net, opt):
|
||||
kw = opt_get(opt_net, ['kwargs'], {})
|
||||
return UResnetMaskProducer(**kw)
|
||||
130
codes/models/vqvae/scaled_weight_conv.py
Normal file
130
codes/models/vqvae/scaled_weight_conv.py
Normal file
|
|
@ -0,0 +1,130 @@
|
|||
from typing import Optional, List
|
||||
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
from torch import Tensor
|
||||
from torch.nn.modules.conv import _ConvNd, _ConvTransposeNd
|
||||
from torch.nn.modules.utils import _ntuple
|
||||
import torch.nn.functional as F
|
||||
|
||||
|
||||
_pair = _ntuple(2)
|
||||
|
||||
|
||||
# Indexes the <p> index of input=b,c,h,w,p by the long tensor index=b,1,h,w. Result is b,c,h,w.
|
||||
# Frankly - IMO - this is what torch.gather should do.
|
||||
def index_2d(input, index):
|
||||
index = index.repeat(1,input.shape[1],1,1)
|
||||
e = torch.eye(input.shape[-1], device=input.device)
|
||||
result = e[index] * input
|
||||
return result.sum(-1)
|
||||
|
||||
|
||||
# Drop-in implementation of Conv2d that can apply masked scales&shifts to the convolution weights.
|
||||
class ScaledWeightConv(_ConvNd):
|
||||
def __init__(
|
||||
self,
|
||||
in_channels: int,
|
||||
out_channels: int,
|
||||
kernel_size,
|
||||
stride = 1,
|
||||
padding = 0,
|
||||
dilation = 1,
|
||||
groups: int = 1,
|
||||
bias: bool = True,
|
||||
padding_mode: str = 'zeros',
|
||||
breadth: int = 8,
|
||||
):
|
||||
stride = _pair(stride)
|
||||
padding = _pair(padding)
|
||||
dilation = _pair(dilation)
|
||||
super().__init__(
|
||||
in_channels, out_channels, _pair(kernel_size), stride, padding, dilation,
|
||||
False, _pair(0), groups, bias, padding_mode)
|
||||
|
||||
self.weight_scales = nn.ParameterList([nn.Parameter(torch.ones(out_channels, in_channels, kernel_size, kernel_size)) for _ in range(breadth)])
|
||||
self.shifts = nn.ParameterList([nn.Parameter(torch.zeros(out_channels, in_channels, kernel_size, kernel_size)) for _ in range(breadth)])
|
||||
for w, s in zip(self.weight_scales, self.shifts):
|
||||
w.FOR_SCALE_SHIFT = True
|
||||
s.FOR_SCALE_SHIFT = True
|
||||
# This should probably be configurable at some point.
|
||||
for p in self.parameters():
|
||||
if not hasattr(p, "FOR_SCALE_SHIFT"):
|
||||
p.DO_NOT_TRAIN = True
|
||||
|
||||
def _weighted_conv_forward(self, input, weight):
|
||||
if self.padding_mode != 'zeros':
|
||||
return F.conv2d(F.pad(input, self._reversed_padding_repeated_twice, mode=self.padding_mode),
|
||||
weight, self.bias, self.stride,
|
||||
_pair(0), self.dilation, self.groups)
|
||||
return F.conv2d(input, weight, self.bias, self.stride,
|
||||
self.padding, self.dilation, self.groups)
|
||||
|
||||
def forward(self, input: Tensor, masks: dict) -> Tensor:
|
||||
# This is an exceptionally inefficient way of achieving this functionality. The hope is that if this is any
|
||||
# good at all, this can be made more efficient by performing a single conv pass with multiple masks.
|
||||
weighted_convs = [self._weighted_conv_forward(input, self.weight * scale + shift) for scale, shift in zip(self.weight_scales, self.shifts)]
|
||||
weighted_convs = torch.stack(weighted_convs, dim=-1)
|
||||
|
||||
needed_mask = weighted_convs.shape[-2]
|
||||
assert needed_mask in masks.keys()
|
||||
|
||||
return index_2d(weighted_convs, masks[needed_mask])
|
||||
|
||||
|
||||
# Drop-in implementation of ConvTranspose2d that can apply masked scales&shifts to the convolution weights.
|
||||
class ScaledWeightConvTranspose(_ConvTransposeNd):
|
||||
def __init__(
|
||||
self,
|
||||
in_channels: int,
|
||||
out_channels: int,
|
||||
kernel_size,
|
||||
stride = 1,
|
||||
padding = 0,
|
||||
output_padding = 0,
|
||||
groups: int = 1,
|
||||
bias: bool = True,
|
||||
dilation: int = 1,
|
||||
padding_mode: str = 'zeros',
|
||||
breadth: int = 8,
|
||||
):
|
||||
stride = _pair(stride)
|
||||
padding = _pair(padding)
|
||||
dilation = _pair(dilation)
|
||||
output_padding = _pair(output_padding)
|
||||
super().__init__(
|
||||
in_channels, out_channels, _pair(kernel_size), stride, padding, dilation,
|
||||
True, output_padding, groups, bias, padding_mode)
|
||||
|
||||
self.weight_scales = nn.ParameterList([nn.Parameter(torch.ones(in_channels, out_channels, kernel_size, kernel_size)) for _ in range(breadth)])
|
||||
self.shifts = nn.ParameterList([nn.Parameter(torch.zeros(in_channels, out_channels, kernel_size, kernel_size)) for _ in range(breadth)])
|
||||
for w, s in zip(self.weight_scales, self.shifts):
|
||||
w.FOR_SCALE_SHIFT = True
|
||||
s.FOR_SCALE_SHIFT = True
|
||||
# This should probably be configurable at some point.
|
||||
for nm, p in self.named_parameters():
|
||||
if nm == 'weight':
|
||||
p.DO_NOT_TRAIN = True
|
||||
|
||||
def _conv_transpose_forward(self, input, weight, output_size) -> Tensor:
|
||||
if self.padding_mode != 'zeros':
|
||||
raise ValueError('Only `zeros` padding mode is supported for ConvTranspose2d')
|
||||
|
||||
output_padding = self._output_padding(
|
||||
input, output_size, self.stride, self.padding, self.kernel_size, self.dilation)
|
||||
|
||||
return F.conv_transpose2d(
|
||||
input, weight, self.bias, self.stride, self.padding,
|
||||
output_padding, self.groups, self.dilation)
|
||||
|
||||
def forward(self, input: Tensor, masks: dict, output_size: Optional[List[int]] = None) -> Tensor:
|
||||
# This is an exceptionally inefficient way of achieving this functionality. The hope is that if this is any
|
||||
# good at all, this can be made more efficient by performing a single conv pass with multiple masks.
|
||||
weighted_convs = [self._conv_transpose_forward(input, self.weight * scale + shift, output_size)
|
||||
for scale, shift in zip(self.weight_scales, self.shifts)]
|
||||
weighted_convs = torch.stack(weighted_convs, dim=-1)
|
||||
|
||||
needed_mask = weighted_convs.shape[-2]
|
||||
assert needed_mask in masks.keys()
|
||||
|
||||
return index_2d(weighted_convs, masks[needed_mask])
|
||||
267
codes/models/vqvae/weighted_conv_vqvae.py
Normal file
267
codes/models/vqvae/weighted_conv_vqvae.py
Normal file
|
|
@ -0,0 +1,267 @@
|
|||
# Copyright 2018 The Sonnet Authors. All Rights Reserved.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
# You may obtain a copy of the License at
|
||||
#
|
||||
# http://www.apache.org/licenses/LICENSE-2.0
|
||||
#
|
||||
# Unless required by applicable law or agreed to in writing, software
|
||||
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
# ============================================================================
|
||||
|
||||
import torch
|
||||
from torch import nn
|
||||
from torch.nn import functional as F
|
||||
|
||||
import torch.distributed as distributed
|
||||
|
||||
from models.vqvae.scaled_weight_conv import ScaledWeightConv, ScaledWeightConvTranspose
|
||||
from trainer.networks import register_model
|
||||
from utils.util import checkpoint, opt_get
|
||||
|
||||
|
||||
class Quantize(nn.Module):
|
||||
def __init__(self, dim, n_embed, decay=0.99, eps=1e-5):
|
||||
super().__init__()
|
||||
|
||||
self.dim = dim
|
||||
self.n_embed = n_embed
|
||||
self.decay = decay
|
||||
self.eps = eps
|
||||
|
||||
embed = torch.randn(dim, n_embed)
|
||||
self.register_buffer("embed", embed)
|
||||
self.register_buffer("cluster_size", torch.zeros(n_embed))
|
||||
self.register_buffer("embed_avg", embed.clone())
|
||||
|
||||
def forward(self, input):
|
||||
flatten = input.reshape(-1, self.dim)
|
||||
dist = (
|
||||
flatten.pow(2).sum(1, keepdim=True)
|
||||
- 2 * flatten @ self.embed
|
||||
+ self.embed.pow(2).sum(0, keepdim=True)
|
||||
)
|
||||
_, embed_ind = (-dist).max(1)
|
||||
embed_onehot = F.one_hot(embed_ind, self.n_embed).type(flatten.dtype)
|
||||
embed_ind = embed_ind.view(*input.shape[:-1])
|
||||
quantize = self.embed_code(embed_ind)
|
||||
|
||||
if self.training:
|
||||
embed_onehot_sum = embed_onehot.sum(0)
|
||||
embed_sum = flatten.transpose(0, 1) @ embed_onehot
|
||||
|
||||
if distributed.is_initialized() and distributed.get_world_size() > 1:
|
||||
distributed.all_reduce(embed_onehot_sum)
|
||||
distributed.all_reduce(embed_sum)
|
||||
|
||||
self.cluster_size.data.mul_(self.decay).add_(
|
||||
embed_onehot_sum, alpha=1 - self.decay
|
||||
)
|
||||
self.embed_avg.data.mul_(self.decay).add_(embed_sum, alpha=1 - self.decay)
|
||||
n = self.cluster_size.sum()
|
||||
cluster_size = (
|
||||
(self.cluster_size + self.eps) / (n + self.n_embed * self.eps) * n
|
||||
)
|
||||
embed_normalized = self.embed_avg / cluster_size.unsqueeze(0)
|
||||
self.embed.data.copy_(embed_normalized)
|
||||
|
||||
diff = (quantize.detach() - input).pow(2).mean()
|
||||
quantize = input + (quantize - input).detach()
|
||||
|
||||
return quantize, diff, embed_ind
|
||||
|
||||
def embed_code(self, embed_id):
|
||||
return F.embedding(embed_id, self.embed.transpose(0, 1))
|
||||
|
||||
|
||||
class ResBlock(nn.Module):
|
||||
def __init__(self, in_channel, channel, breadth):
|
||||
super().__init__()
|
||||
|
||||
self.conv = nn.ModuleList([
|
||||
nn.ReLU(inplace=True),
|
||||
ScaledWeightConv(in_channel, channel, 3, padding=1, breadth=breadth),
|
||||
nn.ReLU(inplace=True),
|
||||
ScaledWeightConv(channel, in_channel, 1, breadth=breadth),
|
||||
])
|
||||
|
||||
def forward(self, input, masks):
|
||||
out = input
|
||||
for m in self.conv:
|
||||
if isinstance(m, ScaledWeightConv):
|
||||
out = m(out, masks)
|
||||
else:
|
||||
out = m(out)
|
||||
out += input
|
||||
|
||||
return out
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
def __init__(self, in_channel, channel, n_res_block, n_res_channel, stride, breadth):
|
||||
super().__init__()
|
||||
|
||||
if stride == 4:
|
||||
blocks = [
|
||||
ScaledWeightConv(in_channel, channel // 2, 4, stride=2, padding=1, breadth=breadth),
|
||||
nn.ReLU(inplace=True),
|
||||
ScaledWeightConv(channel // 2, channel, 4, stride=2, padding=1, breadth=breadth),
|
||||
nn.ReLU(inplace=True),
|
||||
ScaledWeightConv(channel, channel, 3, padding=1, breadth=breadth),
|
||||
]
|
||||
|
||||
elif stride == 2:
|
||||
blocks = [
|
||||
ScaledWeightConv(in_channel, channel // 2, 4, stride=2, padding=1, breadth=breadth),
|
||||
nn.ReLU(inplace=True),
|
||||
ScaledWeightConv(channel // 2, channel, 3, padding=1, breadth=breadth),
|
||||
]
|
||||
|
||||
for i in range(n_res_block):
|
||||
blocks.append(ResBlock(channel, n_res_channel, breadth=breadth))
|
||||
|
||||
blocks.append(nn.ReLU(inplace=True))
|
||||
|
||||
self.blocks = nn.ModuleList(blocks)
|
||||
|
||||
def forward(self, input):
|
||||
for block in self.blocks:
|
||||
if isinstance(block, ScaledWeightConv) or isinstance(block, ResBlock):
|
||||
input = block(input, self.masks)
|
||||
else:
|
||||
input = block(input)
|
||||
return input
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
def __init__(
|
||||
self, in_channel, out_channel, channel, n_res_block, n_res_channel, stride, breadth
|
||||
):
|
||||
super().__init__()
|
||||
|
||||
blocks = [ScaledWeightConv(in_channel, channel, 3, padding=1, breadth=breadth)]
|
||||
|
||||
for i in range(n_res_block):
|
||||
blocks.append(ResBlock(channel, n_res_channel, breadth=breadth))
|
||||
|
||||
blocks.append(nn.ReLU(inplace=True))
|
||||
|
||||
if stride == 4:
|
||||
blocks.extend(
|
||||
[
|
||||
ScaledWeightConvTranspose(channel, channel // 2, 4, stride=2, padding=1, breadth=breadth),
|
||||
nn.ReLU(inplace=True),
|
||||
ScaledWeightConvTranspose(
|
||||
channel // 2, out_channel, 4, stride=2, padding=1, breadth=breadth
|
||||
),
|
||||
]
|
||||
)
|
||||
|
||||
elif stride == 2:
|
||||
blocks.append(
|
||||
ScaledWeightConvTranspose(channel, out_channel, 4, stride=2, padding=1, breadth=breadth)
|
||||
)
|
||||
|
||||
self.blocks = nn.ModuleList(blocks)
|
||||
|
||||
def forward(self, input):
|
||||
for block in self.blocks:
|
||||
if isinstance(block, ScaledWeightConvTranspose) or isinstance(block, ResBlock) \
|
||||
or isinstance(block, ScaledWeightConv):
|
||||
input = block(input, self.masks)
|
||||
else:
|
||||
input = block(input)
|
||||
return input
|
||||
|
||||
|
||||
class VQVAE(nn.Module):
|
||||
def __init__(
|
||||
self,
|
||||
in_channel=3,
|
||||
channel=128,
|
||||
n_res_block=2,
|
||||
n_res_channel=32,
|
||||
codebook_dim=64,
|
||||
codebook_size=512,
|
||||
breadth=8,
|
||||
decay=0.99,
|
||||
):
|
||||
super().__init__()
|
||||
|
||||
self.enc_b = Encoder(in_channel, channel, n_res_block, n_res_channel, stride=4, breadth=breadth)
|
||||
self.enc_t = Encoder(channel, channel, n_res_block, n_res_channel, stride=2, breadth=breadth)
|
||||
self.quantize_conv_t = ScaledWeightConv(channel, codebook_dim, 1, breadth=breadth)
|
||||
self.quantize_t = Quantize(codebook_dim, codebook_size)
|
||||
self.dec_t = Decoder(
|
||||
codebook_dim, codebook_dim, channel, n_res_block, n_res_channel, stride=2, breadth=breadth
|
||||
)
|
||||
self.quantize_conv_b = ScaledWeightConv(codebook_dim + channel, codebook_dim, 1, breadth=breadth)
|
||||
self.quantize_b = Quantize(codebook_dim, codebook_size)
|
||||
self.upsample_t = ScaledWeightConvTranspose(
|
||||
codebook_dim, codebook_dim, 4, stride=2, padding=1, breadth=breadth
|
||||
)
|
||||
self.dec = Decoder(
|
||||
codebook_dim + codebook_dim,
|
||||
in_channel,
|
||||
channel,
|
||||
n_res_block,
|
||||
n_res_channel,
|
||||
stride=4,
|
||||
breadth=breadth
|
||||
)
|
||||
|
||||
def forward(self, input, masks):
|
||||
# This awkward injection point is necessary to enable checkpointing to work.
|
||||
for m in [self.enc_b, self.enc_t, self.dec_t, self.dec]:
|
||||
m.masks = masks
|
||||
|
||||
quant_t, quant_b, diff, _, _ = self.encode(input, masks)
|
||||
dec = self.decode(quant_t, quant_b, masks)
|
||||
|
||||
return dec, diff
|
||||
|
||||
def encode(self, input, masks):
|
||||
enc_b = checkpoint(self.enc_b, input)
|
||||
enc_t = checkpoint(self.enc_t, enc_b)
|
||||
|
||||
quant_t = self.quantize_conv_t(enc_t, masks).permute(0, 2, 3, 1)
|
||||
quant_t, diff_t, id_t = self.quantize_t(quant_t)
|
||||
quant_t = quant_t.permute(0, 3, 1, 2)
|
||||
diff_t = diff_t.unsqueeze(0)
|
||||
|
||||
dec_t = checkpoint(self.dec_t, quant_t)
|
||||
enc_b = torch.cat([dec_t, enc_b], 1)
|
||||
|
||||
quant_b = self.quantize_conv_b(enc_b, masks).permute(0, 2, 3, 1)
|
||||
quant_b, diff_b, id_b = self.quantize_b(quant_b)
|
||||
quant_b = quant_b.permute(0, 3, 1, 2)
|
||||
diff_b = diff_b.unsqueeze(0)
|
||||
|
||||
return quant_t, quant_b, diff_t + diff_b, id_t, id_b
|
||||
|
||||
def decode(self, quant_t, quant_b, masks):
|
||||
upsample_t = self.upsample_t(quant_t, masks)
|
||||
quant = torch.cat([upsample_t, quant_b], 1)
|
||||
dec = checkpoint(self.dec, quant)
|
||||
|
||||
return dec
|
||||
|
||||
def decode_code(self, code_t, code_b):
|
||||
quant_t = self.quantize_t.embed_code(code_t)
|
||||
quant_t = quant_t.permute(0, 3, 1, 2)
|
||||
quant_b = self.quantize_b.embed_code(code_b)
|
||||
quant_b = quant_b.permute(0, 3, 1, 2)
|
||||
|
||||
dec = self.decode(quant_t, quant_b, masks)
|
||||
|
||||
return dec
|
||||
|
||||
|
||||
@register_model
|
||||
def register_weighted_vqvae(opt_net, opt):
|
||||
kw = opt_get(opt_net, ['kwargs'], {})
|
||||
return VQVAE(**kw)
|
||||
|
|
@ -295,7 +295,7 @@ class Trainer:
|
|||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument('-opt', type=str, help='Path to option YAML file.', default='../options/train_imgset_stylesr.yml')
|
||||
parser.add_argument('-opt', type=str, help='Path to option YAML file.', default='../experiments/train_imgset_vqvae_stage1/train_imgset_vqvae_stage1_5.yml')
|
||||
parser.add_argument('--launcher', choices=['none', 'pytorch'], default='none', help='job launcher')
|
||||
parser.add_argument('--local_rank', type=int, default=0)
|
||||
args = parser.parse_args()
|
||||
|
|
|
|||
|
|
@ -115,8 +115,7 @@ def kmeans(
|
|||
def kmeans_predict(
|
||||
X,
|
||||
cluster_centers,
|
||||
distance='euclidean',
|
||||
device=torch.device('cpu')
|
||||
distance='euclidean'
|
||||
):
|
||||
"""
|
||||
predict using cluster centers
|
||||
|
|
@ -126,8 +125,6 @@ def kmeans_predict(
|
|||
:param device: (torch.device) device [default: 'cpu']
|
||||
:return: (torch.tensor) cluster ids
|
||||
"""
|
||||
print(f'predicting on {device}..')
|
||||
|
||||
if distance == 'euclidean':
|
||||
pairwise_distance_function = pairwise_distance
|
||||
elif distance == 'cosine':
|
||||
|
|
@ -135,22 +132,13 @@ def kmeans_predict(
|
|||
else:
|
||||
raise NotImplementedError
|
||||
|
||||
# convert to float
|
||||
X = X.float()
|
||||
|
||||
# transfer to device
|
||||
X = X.to(device)
|
||||
|
||||
dis = pairwise_distance_function(X, cluster_centers)
|
||||
choice_cluster = torch.argmin(dis, dim=1)
|
||||
|
||||
return choice_cluster.cpu()
|
||||
return choice_cluster
|
||||
|
||||
|
||||
def pairwise_distance(data1, data2, device=torch.device('cpu')):
|
||||
# transfer to device
|
||||
data1, data2 = data1.to(device), data2.to(device)
|
||||
|
||||
def pairwise_distance(data1, data2):
|
||||
# N*1*M
|
||||
A = data1.unsqueeze(dim=1)
|
||||
|
||||
|
|
@ -163,10 +151,7 @@ def pairwise_distance(data1, data2, device=torch.device('cpu')):
|
|||
return dis
|
||||
|
||||
|
||||
def pairwise_cosine(data1, data2, device=torch.device('cpu')):
|
||||
# transfer to device
|
||||
data1, data2 = data1.to(device), data2.to(device)
|
||||
|
||||
def pairwise_cosine(data1, data2):
|
||||
# N*1*M
|
||||
A = data1.unsqueeze(dim=1)
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue
Block a user