Get rid of SwitchedResidualGenerator
Just use the configurable one instead..
This commit is contained in:
James Betker
parent
379b96eb55
commit
7d541642aa
|
|
@ -86,84 +86,13 @@ class SwitchComputer(nn.Module):
|
|||
multiplexer = self.proc_switch_conv(multiplexer)
|
||||
multiplexer = self.final_switch_conv.forward(multiplexer)
|
||||
# Interpolate the multiplexer across the entire shape of the image.
|
||||
multiplexer = F.interpolate(multiplexer, size=x.shape[2:], mode='nearest')
|
||||
multiplexer = F.interpolate(multiplexer, size=x.shape[2:], mode='nearest', recompute_scale_factor=True)
|
||||
|
||||
return self.switch(xformed, multiplexer, output_attention_weights)
|
||||
|
||||
def set_temperature(self, temp):
|
||||
self.switch.set_attention_temperature(temp)
|
||||
|
||||
class SwitchedResidualGenerator(nn.Module):
|
||||
def __init__(self, switch_filters, initial_temp=20, final_temperature_step=50000):
|
||||
super(SwitchedResidualGenerator, self).__init__()
|
||||
self.switch1 = SwitchComputer(3, switch_filters, functools.partial(ResidualBranch, 3, 3, kernel_size=7, depth=3), 4, 4, 0, initial_temp)
|
||||
self.switch2 = SwitchComputer(3, switch_filters, functools.partial(ResidualBranch, 3, 3, kernel_size=5, depth=3), 8, 3, 0, initial_temp)
|
||||
self.switch3 = SwitchComputer(3, switch_filters, functools.partial(ResidualBranch, 3, 3, kernel_size=3, depth=3), 16, 2, 1, initial_temp)
|
||||
self.switch4 = SwitchComputer(3, switch_filters * 2, functools.partial(ResidualBranch, 3, 3, kernel_size=3, depth=2), 32, 1, 2, initial_temp)
|
||||
initialize_weights([self.switch1, self.switch2, self.switch3, self.switch4], 1)
|
||||
# Initialize the transforms with a lesser weight, since they are repeatedly added on to the resultant image.
|
||||
initialize_weights([self.switch1.transforms, self.switch2.transforms, self.switch3.transforms, self.switch4.transforms], .05)
|
||||
|
||||
self.init_temperature = initial_temp
|
||||
self.final_temperature_step = final_temperature_step
|
||||
self.running_sum = [0, 0, 0, 0]
|
||||
self.running_hist = [[],[],[],[]]
|
||||
self.running_count = 0
|
||||
|
||||
def forward(self, x):
|
||||
sw1, self.a1 = self.switch1.forward(x, True)
|
||||
x = x + sw1
|
||||
sw2, self.a2 = self.switch2.forward(x, True)
|
||||
x = x + sw2
|
||||
sw3, self.a3 = self.switch3.forward(x, True)
|
||||
x = x + sw3
|
||||
sw4, self.a4 = self.switch4.forward(x, True)
|
||||
x = x + sw4
|
||||
|
||||
a1mean, a1i = compute_attention_specificity(self.a1, 2)
|
||||
a2mean, a2i = compute_attention_specificity(self.a2, 2)
|
||||
a3mean, a3i = compute_attention_specificity(self.a3, 2)
|
||||
a4mean, a4i = compute_attention_specificity(self.a4, 2)
|
||||
running_sum = [
|
||||
self.running_sum[0] + a1mean,
|
||||
self.running_sum[1] + a2mean,
|
||||
self.running_sum[2] + a3mean,
|
||||
self.running_sum[3] + a4mean,
|
||||
]
|
||||
self.running_hist[0].append(a1i.detach().cpu().flatten())
|
||||
self.running_hist[1].append(a2i.detach().cpu().flatten())
|
||||
self.running_hist[2].append(a3i.detach().cpu().flatten())
|
||||
self.running_hist[3].append(a4i.detach().cpu().flatten())
|
||||
self.running_count += 1
|
||||
|
||||
return (x,)
|
||||
|
||||
def set_temperature(self, temp):
|
||||
self.switch1.set_temperature(temp)
|
||||
self.switch2.set_temperature(temp)
|
||||
self.switch3.set_temperature(temp)
|
||||
self.switch4.set_temperature(temp)
|
||||
|
||||
def get_debug_values(self, step):
|
||||
# Take the chance to update the temperature here.
|
||||
temp = max(1, int(self.init_temperature * (self.final_temperature_step - step) / self.final_temperature_step))
|
||||
self.set_temperature(temp)
|
||||
|
||||
if step % 250 == 0:
|
||||
save_attention_to_image(self.a1, 4, step, "a1")
|
||||
save_attention_to_image(self.a2, 8, step, "a2", 2)
|
||||
save_attention_to_image(self.a3, 16, step, "a3", 4)
|
||||
save_attention_to_image(self.a4, 32, step, "a4", 8)
|
||||
|
||||
val = {"switch_temperature": temp}
|
||||
for i in range(len(self.running_sum)):
|
||||
val["switch_%i_specificity" % (i,)] = self.running_sum[i] / self.running_count
|
||||
val["switch_%i_histogram" % (i,)] = torch.cat(self.running_hist[i])
|
||||
self.running_sum[i] = 0
|
||||
self.running_hist[i] = []
|
||||
self.running_count = 0
|
||||
return val
|
||||
|
||||
|
||||
class ConfigurableSwitchedResidualGenerator(nn.Module):
|
||||
def __init__(self, switch_filters, switch_reductions, switch_processing_layers, trans_counts, trans_kernel_sizes, trans_layers, initial_temp=20, final_temperature_step=50000):
|
||||
|
|
@ -178,22 +107,13 @@ class ConfigurableSwitchedResidualGenerator(nn.Module):
|
|||
self.transformation_counts = trans_counts
|
||||
self.init_temperature = initial_temp
|
||||
self.final_temperature_step = final_temperature_step
|
||||
self.running_sum = [0 for i in range(len(switches))]
|
||||
self.running_hist = [[] for i in range(len(switches))]
|
||||
self.running_count = 0
|
||||
|
||||
def forward(self, x):
|
||||
self.attentions = []
|
||||
for i, sw in enumerate(self.switches):
|
||||
x, att = sw.forward(x, True)
|
||||
self.attentions.append(att)
|
||||
spec, hist = compute_attention_specificity(att, 2)
|
||||
self.running_sum[i] += spec
|
||||
self.running_hist[i].append(hist.detach().cpu().flatten())
|
||||
|
||||
self.running_count += 1
|
||||
|
||||
return (x,)
|
||||
return x,
|
||||
|
||||
def set_temperature(self, temp):
|
||||
[sw.set_temperature(temp) for sw in self.switches]
|
||||
|
|
@ -206,11 +126,11 @@ class ConfigurableSwitchedResidualGenerator(nn.Module):
|
|||
if step % 250 == 0:
|
||||
[save_attention_to_image(self.attentions[i], self.transformation_counts[i], step, "a%i" % (i+1,), l_mult=float(self.transformation_counts[i]/4)) for i in range(len(self.switches))]
|
||||
|
||||
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(self.running_sum)):
|
||||
val["switch_%i_specificity" % (i,)] = self.running_sum[i] / self.running_count
|
||||
self.running_sum[i] = 0
|
||||
val["switch_%i_histogram" % (i,)] = torch.cat(self.running_hist[i])
|
||||
self.running_hist[i] = []
|
||||
self.running_count = 0
|
||||
for i in range(len(means)):
|
||||
val["switch_%i_specificity" % (i,)] = means[i]
|
||||
val["switch_%i_histogram" % (i,)] = hists[i]
|
||||
return val
|
||||
|
|
@ -62,16 +62,6 @@ def define_G(opt, net_key='network_G'):
|
|||
init_temperature=opt_net['temperature'],
|
||||
final_temperature_step=opt_net['temperature_final_step'])
|
||||
netG = RRDBNet_arch.PixShuffleRRDB(nf=opt_net['nf'], nb=opt_net['nb'], gc=opt_net['gc'], scale=scale, rrdb_block_f=block_f)
|
||||
elif which_model == 'ResGen':
|
||||
netG = ResGen_arch.fixup_resnet34(nb_denoiser=opt_net['nb_denoiser'], nb_upsampler=opt_net['nb_upsampler'],
|
||||
upscale_applications=opt_net['upscale_applications'], num_filters=opt_net['nf'])
|
||||
elif which_model == 'ResGenV2':
|
||||
netG = ResGen_arch.fixup_resnet34_v2(nb_denoiser=opt_net['nb_denoiser'], nb_upsampler=opt_net['nb_upsampler'],
|
||||
upscale_applications=opt_net['upscale_applications'], num_filters=opt_net['nf'],
|
||||
inject_noise=opt_net['inject_noise'])
|
||||
elif which_model == "SwitchedResidualGenerator":
|
||||
netG = SwitchedGen_arch.SwitchedResidualGenerator(switch_filters=opt_net['nf'], initial_temp=opt_net['temperature'],
|
||||
final_temperature_step=opt_net['temperature_final_step'])
|
||||
elif which_model == "ConfigurableSwitchedResidualGenerator":
|
||||
netG = SwitchedGen_arch.ConfigurableSwitchedResidualGenerator(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'],
|
||||
|
|
|
|||
Loading…
Reference in New Issue
Block a user