Step 2：预处理

Step 1：创建自定义数据集

w = torch.rand(16, 3, 5, 5)

= (ker_num, input_channel, ker_size, ker_size)

Input_channels:

• 黑白：1
• 彩色：3

Stocastic: 随即筛选样本

val_set: for detecting overfitting

torch.nn.function

.matmul() 取后两维相乘

unsqueeze:

正：在之前插入

负：在之后插入

.index_select(0, [0, 2])

torch.tensor([2., 3.2])

torch.FloatTensor(2, 3)

Unintialized: 未初始化的tensor

增强学习一般用 DoubleTensor

Likelihood

Sequence Generation

Conditional Sequence Generation

Maximizing Expected Reward

Policy Gradient

Conditional GAN

Abtractive Summarization

GAN + Autoencoder

Feature Extraction:

InfoGAN

VAE-GAN

BiGAN

Triple GAN

Feature Disentangle  v. 解开

J-S divergence proplem

Wasserstein GAN:

Earth Mover's Distance

Lipschitz Function

intractable adj. 棘手的 <==> difficult

f-divergence

exponential  adj. 指数

Theory behind GAN:

Divergence

KL Divergence

sample v. 抽样

J-S divergence

Unsupervised Conditional Generation

• Direct Transfomation
• Projection to Common Space:

CycleGAN:

Cycle consistency

GAN: Generative Adversarial Network

since sliced bread

Disciminator

Step 1: Fix G, update D

Step 1: Fix D, update G

Can Generator learn by itself?

Auto-encoder

Decoder = Generator

Can Discriminator generate?