generative-adversarial-network

问题 It's the first time I'm working with GANs and I am facing an issue regarding the Discriminator repeatedly outperforming the Generator. I am trying to reproduce the PA model from this article and I'm looking at this slightly different implementation to help me out. I have read quite a lot of papers on how GANs work and also followed some tutorials to understand them better. Moreover, I've read articles on how to overcome the major instabilities, but I can't find a way to overcome this behavior

问题 Trying combine GAN generator and critic to train both as VAE. Base code is here. Code modified to create encoder on top of critic: def _build_critic(self): #### THE critic critic_input = Input(shape=self.input_dim, name='critic_input') x = critic_input for i in range(self.n_layers_critic): x = Conv2D( filters = self.critic_conv_filters[i] , kernel_size = self.critic_conv_kernel_size[i] , strides = self.critic_conv_strides[i] , padding = 'same' , name = 'critic_conv_' + str(i) , kernel

问题 I have trained a GAN to reproduce CIFAR10 like images. Initially I notice all images cross one batch produced by the generator look always the same, like the picture below: After hours of debugging and comparison to the tutorial which is a great learning source for beginners (https://machinelearningmastery.com/how-to-develop-a-generative-adversarial-network-for-a-cifar-10-small-object-photographs-from-scratch/), I just add only one letter on my original code and the generated images start

问题 I have trained a GAN to reproduce CIFAR10 like images. Initially I notice all images cross one batch produced by the generator look always the same, like the picture below: After hours of debugging and comparison to the tutorial which is a great learning source for beginners (https://machinelearningmastery.com/how-to-develop-a-generative-adversarial-network-for-a-cifar-10-small-object-photographs-from-scratch/), I just add only one letter on my original code and the generated images start

问题 I have trained a GAN to reproduce CIFAR10 like images. Initially I notice all images cross one batch produced by the generator look always the same, like the picture below: After hours of debugging and comparison to the tutorial which is a great learning source for beginners (https://machinelearningmastery.com/how-to-develop-a-generative-adversarial-network-for-a-cifar-10-small-object-photographs-from-scratch/), I just add only one letter on my original code and the generated images start

问题 So here's whats happening. I have been using the StyleGAN2 model for a while now and I decided to make a website that will allow the user to input the arguments for the model to generate the images. The model has been trained using tensorflow v1.15 and the code works perfectly fine and generates all the required outputs when I run the model directly on my machine through the command line. The problem arises when I am now using a flask API to do the same thing. Here is all the code for

问题 So here's whats happening. I have been using the StyleGAN2 model for a while now and I decided to make a website that will allow the user to input the arguments for the model to generate the images. The model has been trained using tensorflow v1.15 and the code works perfectly fine and generates all the required outputs when I run the model directly on my machine through the command line. The problem arises when I am now using a flask API to do the same thing. Here is all the code for

问题 I am working on an OCR system. A challenge that I'm facing for recognizing the text within ROI is due to the shakiness or motion effect shot or text that is not focus due to angle positions . Please consider the following demo sample If you notice the texts (for ex. the mark as a red), in such cases the OCR system couldn't properly recognize the text. However, this scenario can also come on with no angle shot where the image is too blurry that the OCR system can't recognize or partially

问题 I am trying to implement StyleGAN with tensorflow version 2, and i have no ideas how to do equalized learning rate. I tried to scale gradients with this way: gradeints equalization But it doesn't works correctly. Pls help. 回答1: You can just create a custom layer. class DenseEQ(Dense): """ Standard dense layer but includes learning rate equilization at runtime as per Karras et al. 2017. Inherits Dense layer and overides the call method. """ def __init__(self, **kwargs): if 'kernel_initializer'

问题 I'm designing a custom layer for my neural network, but I get an error from my code. I want to do a attention layer as described in the paper: SAGAN. And the original tf code class AttentionLayer(Layer): def __init__(self, **kwargs): super(AttentionLayer, self).__init__(**kwargs) def build(self, input_shape): input_dim = input_shape[-1] filters_f_g = input_dim // 8 filters_h = input_dim kernel_shape_f_g = (1, 1) + (input_dim, filters_f_g) kernel_shape_h = (1, 1) + (input_dim, filters_h) #