# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve. # # 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 paddle from paddle import nn from paddle.nn import functional as F from .builder import GENERATORS from ...modules.conv import ConvBNRelu from ...modules.conv import NonNormConv2d from ...modules.conv import Conv2dTransposeRelu @GENERATORS.register() class Wav2Lip(nn.Layer): def __init__(self): super(Wav2Lip, self).__init__() self.face_encoder_blocks = [ nn.Sequential(ConvBNRelu(6, 16, kernel_size=7, stride=1, padding=3)), # 96,96 nn.Sequential( ConvBNRelu(16, 32, kernel_size=3, stride=2, padding=1), # 48,48 ConvBNRelu(32, 32, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(32, 32, kernel_size=3, stride=1, padding=1, residual=True)), nn.Sequential( ConvBNRelu(32, 64, kernel_size=3, stride=2, padding=1), # 24,24 ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True)), nn.Sequential( ConvBNRelu(64, 128, kernel_size=3, stride=2, padding=1), # 12,12 ConvBNRelu(128, 128, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(128, 128, kernel_size=3, stride=1, padding=1, residual=True)), nn.Sequential( ConvBNRelu(128, 256, kernel_size=3, stride=2, padding=1), # 6,6 ConvBNRelu(256, 256, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(256, 256, kernel_size=3, stride=1, padding=1, residual=True)), nn.Sequential( ConvBNRelu(256, 512, kernel_size=3, stride=2, padding=1), # 3,3 ConvBNRelu(512, 512, kernel_size=3, stride=1, padding=1, residual=True), ), nn.Sequential( ConvBNRelu(512, 512, kernel_size=3, stride=1, padding=0), # 1, 1 ConvBNRelu(512, 512, kernel_size=1, stride=1, padding=0)), ] self.audio_encoder = nn.Sequential( ConvBNRelu(1, 32, kernel_size=3, stride=1, padding=1), ConvBNRelu(32, 32, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(32, 32, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(32, 64, kernel_size=3, stride=(3, 1), padding=1), ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(64, 128, kernel_size=3, stride=3, padding=1), ConvBNRelu(128, 128, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(128, 128, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(128, 256, kernel_size=3, stride=(3, 2), padding=1), ConvBNRelu(256, 256, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(256, 512, kernel_size=3, stride=1, padding=0), ConvBNRelu(512, 512, kernel_size=1, stride=1, padding=0), ) self.face_decoder_blocks = [ nn.Sequential( ConvBNRelu(512, 512, kernel_size=1, stride=1, padding=0), ), nn.Sequential( Conv2dTransposeRelu(1024, 512, kernel_size=3, stride=1, padding=0), # 3,3 ConvBNRelu(512, 512, kernel_size=3, stride=1, padding=1, residual=True), ), nn.Sequential( Conv2dTransposeRelu(1024, 512, kernel_size=3, stride=2, padding=1, output_padding=1), ConvBNRelu(512, 512, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(512, 512, kernel_size=3, stride=1, padding=1, residual=True), ), # 6, 6 nn.Sequential( Conv2dTransposeRelu(768, 384, kernel_size=3, stride=2, padding=1, output_padding=1), ConvBNRelu(384, 384, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(384, 384, kernel_size=3, stride=1, padding=1, residual=True), ), # 12, 12 nn.Sequential( Conv2dTransposeRelu(512, 256, kernel_size=3, stride=2, padding=1, output_padding=1), ConvBNRelu(256, 256, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(256, 256, kernel_size=3, stride=1, padding=1, residual=True), ), # 24, 24 nn.Sequential( Conv2dTransposeRelu(320, 128, kernel_size=3, stride=2, padding=1, output_padding=1), ConvBNRelu(128, 128, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(128, 128, kernel_size=3, stride=1, padding=1, residual=True), ), # 48, 48 nn.Sequential( Conv2dTransposeRelu(160, 64, kernel_size=3, stride=2, padding=1, output_padding=1), ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True), ConvBNRelu(64, 64, kernel_size=3, stride=1, padding=1, residual=True), ), ] # 96,96 self.output_block = nn.Sequential( ConvBNRelu(80, 32, kernel_size=3, stride=1, padding=1), nn.Conv2D(32, 3, kernel_size=1, stride=1, padding=0), nn.Sigmoid()) def forward(self, audio_sequences, face_sequences): # audio_sequences = (B, T, 1, 80, 16) B = audio_sequences.shape[0] input_dim_size = len(face_sequences.shape) if input_dim_size > 4: audio_sequences = paddle.concat([ audio_sequences[:, i] for i in range(audio_sequences.shape[1]) ], axis=0) face_sequences = paddle.concat([ face_sequences[:, :, i] for i in range(face_sequences.shape[2]) ], axis=0) audio_embedding = self.audio_encoder(audio_sequences) # B, 512, 1, 1 feats = [] x = face_sequences for f in self.face_encoder_blocks: x = f(x) feats.append(x) x = audio_embedding for f in self.face_decoder_blocks: x = f(x) try: x = paddle.concat((x, feats[-1]), axis=1) except Exception as e: print(x.shape) print(feats[-1].shape) raise e feats.pop() x = self.output_block(x) if input_dim_size > 4: x = paddle.split(x, B, axis=0) # [(B, C, H, W)] outputs = paddle.stack(x, axis=2) # (B, C, T, H, W) else: outputs = x return outputs class Wav2LipDiscQual(nn.Layer): def __init__(self): super(Wav2LipDiscQual, self).__init__() self.face_encoder_blocks = [ nn.Sequential( NonNormConv2d(3, 32, kernel_size=7, stride=1, padding=3)), # 48,96 nn.Sequential( NonNormConv2d(32, 64, kernel_size=5, stride=(1, 2), padding=2), # 48,48 NonNormConv2d(64, 64, kernel_size=5, stride=1, padding=2)), nn.Sequential( NonNormConv2d(64, 128, kernel_size=5, stride=2, padding=2), # 24,24 NonNormConv2d(128, 128, kernel_size=5, stride=1, padding=2)), nn.Sequential( NonNormConv2d(128, 256, kernel_size=5, stride=2, padding=2), # 12,12 NonNormConv2d(256, 256, kernel_size=5, stride=1, padding=2)), nn.Sequential( NonNormConv2d(256, 512, kernel_size=3, stride=2, padding=1), # 6,6 NonNormConv2d(512, 512, kernel_size=3, stride=1, padding=1)), nn.Sequential( NonNormConv2d(512, 512, kernel_size=3, stride=2, padding=1), # 3,3 NonNormConv2d(512, 512, kernel_size=3, stride=1, padding=1), ), nn.Sequential( NonNormConv2d(512, 512, kernel_size=3, stride=1, padding=0), # 1, 1 NonNormConv2d(512, 512, kernel_size=1, stride=1, padding=0)), ] self.binary_pred = nn.Sequential( nn.Conv2D(512, 1, kernel_size=1, stride=1, padding=0), nn.Sigmoid()) self.label_noise = .0 def get_lower_half(self, face_sequences): return face_sequences[:, :, face_sequences.shape[2] // 2:] def to_2d(self, face_sequences): B = face_sequences.shape[0] face_sequences = paddle.concat( [face_sequences[:, :, i] for i in range(face_sequences.shape[2])], axis=0) return face_sequences def perceptual_forward(self, false_face_sequences): false_face_sequences = self.to_2d(false_face_sequences) false_face_sequences = self.get_lower_half(false_face_sequences) false_feats = false_face_sequences for f in self.face_encoder_blocks: false_feats = f(false_feats) false_pred_loss = F.binary_cross_entropy( paddle.reshape(self.binary_pred(false_feats), (len(false_feats), -1)), paddle.ones((len(false_feats), 1))) return false_pred_loss def forward(self, face_sequences): face_sequences = self.to_2d(face_sequences) face_sequences = self.get_lower_half(face_sequences) x = face_sequences for f in self.face_encoder_blocks: x = f(x) return paddle.reshape(self.binary_pred(x), (len(x), -1))