Questions about visual feature encoder

[dengyuanjie]

Could you please provide the code for the training process of the visual feature encoder? That is, this part of the audio-visual joint representation learning and adversarial training described in III.A of the paper. Thanks！

At the same time, I did not find the pre-trained model in the following path. I only found the mobilenet0.25_Final.pth file.
parser.add_argument('--trained_visual_encoder', default='./weights/visual_feature_encoder.pth', type=str, help='Trained state_dict file path to open')