def __init__(self, text_encoder, config):
super(TextImage, self).__init__()
self.config = config
self.TextEncoderBottom = text_encoder
self.TextEncoderTop = TextEncoderTop(**config['TextEncoderTop'])
self.ImageEncoder = ImageEncoder(**config['ImageEncoder'])
self.optimizer = optim.Adam(self.parameters(), lr=config['lr'])
def __init__(self, speech_encoder, config):
super(SpeechCorrText, self).__init__()
self.config = config
self.SpeechEncoderBottom = speech_encoder
self.SpeechEncoderTop = SpeechEncoderTop(**config['SpeechEncoderTop'])
self.CorrEncoder = ImageEncoder(size=1024, size_target=1024 * 4)
self.optimizer = optim.Adam(self.parameters(), lr=config['lr'])
def __init__(self, speech_encoder, config):
super(SpeechImage, self).__init__()
self.config = config
self.SpeechEncoderBottom = speech_encoder
self.SpeechEncoderTop = SpeechEncoderTop(**config['SpeechEncoderTop'])
self.ImageEncoder = ImageEncoder(**config['ImageEncoder'])
self.optimizer = optim.Adam(self.parameters(), lr=config['lr'])
