Remove UnetPro, make float values visible in README

face_swapper/config.ini

@@ -16,7 +16,6 @@ gazer_path =
 motion_extractor_path =
 
 [training.model.generator]
-encoder_type =
 identity_channels =
 output_channels =
 output_size =

face_swapper/src/models/generator.py

@@ -3,7 +3,7 @@ import configparser
 from torch import Tensor, nn
 
 from ..networks.aad import AAD
-from ..networks.unet import UNet, UNetPro
+from ..networks.unet import UNet
 from ..types import Attributes, Embedding
 
 CONFIG = configparser.ConfigParser()
@@ -13,16 +13,12 @@ CONFIG.read('config.ini')
 class Generator(nn.Module):
 	def __init__(self) -> None:
 		super().__init__()
-		encoder_type = CONFIG.get('training.model.generator', 'encoder_type')
 		identity_channels = CONFIG.getint('training.model.generator', 'identity_channels')
 		output_channels = CONFIG.getint('training.model.generator', 'output_channels')
 		output_size = CONFIG.getint('training.model.generator', 'output_size')
 		num_blocks = CONFIG.getint('training.model.generator', 'num_blocks')
 
-		if encoder_type == 'unet':
-			self.encoder = UNet(output_size)
-		if encoder_type == 'unet-pro':
-			self.encoder = UNetPro(output_size)
+		self.encoder = UNet(output_size)
 		self.generator = AAD(identity_channels, output_channels, output_size, num_blocks)
 		self.encoder.apply(init_weight)
 		self.generator.apply(init_weight)

face_swapper/src/models/loss.py

@@ -169,7 +169,7 @@ class GazeLoss(nn.Module):
 		return gaze_loss, weighted_gaze_loss
 
 	def detect_gaze(self, input_tensor : Tensor) -> Gaze:
-		scale_factor = CONFIG.getint('training.losses', 'gaze_scale_factor')
+		scale_factor = CONFIG.getfloat('training.losses', 'gaze_scale_factor')
 		y_min = int(60 * scale_factor)
 		y_max = int(224 * scale_factor)
 		x_min = int(16 * scale_factor)

face_swapper/src/networks/unet.py

@@ -2,9 +2,6 @@ from typing import Tuple
 
 import torch
 from torch import Tensor, nn
-from torchvision import models
-from torchvision.models import ResNet34_Weights
-
 
 class UNet(nn.Module):
 	def __init__(self, output_size : int) -> None:
@@ -88,38 +85,6 @@ class UNet(nn.Module):
 		return bottleneck_tensor, *up_features, output_tensor
 
 
-class UNetPro(UNet):
-	def __init__(self, output_size : int) -> None:
-		super().__init__(output_size)
-		self.resnet = models.resnet34(weights = ResNet34_Weights.DEFAULT)
-		self.down_samples = self.create_down_samples()
-		self.up_samples = self.create_up_samples()
-
-	def create_down_samples(self) -> nn.ModuleList:
-		down_samples = nn.ModuleList(
-		[
-			nn.Sequential(
-				self.resnet.conv1,
-				self.resnet.bn1,
-				self.resnet.relu,
-				nn.Conv2d(64, 32, kernel_size = 1, bias = False),
-				nn.BatchNorm2d(32),
-				nn.LeakyReLU(0.1, inplace = True)
-			),
-			DownSample(32, 64),
-			self.resnet.layer2,
-			self.resnet.layer3,
-			self.resnet.layer4,
-			DownSample(512, 1024),
-			DownSample(1024, 1024)
-		])
-
-		if self.output_size == 512:
-			down_samples.append(DownSample(2048, 2048))
-
-		return down_samples
-
-
 class UpSample(nn.Module):
 	def __init__(self, input_channels : int, output_channels : int) -> None:
 		super().__init__()

face_swapper/src/training.py

@@ -80,7 +80,7 @@ class FaceSwapperTrainer(lightning.LightningModule):
 		return generator_config, discriminator_config
 
 	def training_step(self, batch : Batch, batch_index : int) -> Tensor:
-		preview_frequency = CONFIG.getfloat('training.trainer', 'preview_frequency')
+		preview_frequency = CONFIG.getint('training.trainer', 'preview_frequency')
 
 		source_tensor, target_tensor = batch
 		generator_optimizer, discriminator_optimizer = self.optimizers() #type:ignore[attr-defined]