WGAN-GP實戰

在實現了GAN之後，下面就來看一下WGAN-GP實戰

看一下WGAN如何解決training不穩定的問題

 

加了個1.3 gradientr penalty函數

 

 

wgan.py

import torch
from torch import nn, optim, autograd
import numpy as np
import visdom
import random
from matplotlib import pyplot as plt

h_dim = 400
batchsz = 512
viz = visdom.Visdom()

class Generator(nn.Module):
	def __init__(self):
		super(Generator, self).__init__()

		#z:[b.2] => [b,2]
		self.net = nn.Sequential(    #一共4層
			nn.Linear(2, h_dim),
			nn.ReLU(True),
			nn.Linear(h_dim, h_dim),
			nn.ReLU(True),
			nn.Linear(h_dim, h_dim),
			nn.ReLU(True),
			nn.Linear(h_dim, 2),

			)

	def forward(self, z):
		output = self.net(z)
		return output



class Discriminator(nn.Module):

	def __init__(self):
		super(Discriminator, self).__init__()

		self.net = nn.Sequential( #4層
			nn.Linear(2, h_dim),
			nn.ReLU(True),
			nn.Linear(h_dim, h_dim),
			nn.ReLU(True),
			nn.Linear(h_dim, h_dim),
			nn.ReLU(True),
			nn.Linear(h_dim, 1),
			nn.Sigmoid()   #因爲discriminator的輸入是probability，用sigmoid函數把它弄到(0,1)的範圍

			)


	def forward(self, x):
		output = self.net(x)
		return output.view(-1)


def data_generator():
	#8-gaussian mixture models

	scale = 2.
	centers = [
		(1, 0),
		(-1, 0),
		(0, 1),
		(0, -1),
		(1. / np.sqrt(2), 1. / np.sqrt(2)),
		(1. / np.sqrt(2), -1. / np.sqrt(2)),
		(-1. / np.sqrt(2), 1. / np.sqrt(2)),
		(-1. / np.sqrt(2), -1. / np.sqrt(2))
	]
	centers = [(scale * x, scale * y) for x,y in centers]

	while True:
		dataset = []
		for i in range(batchsz):
			point = np.random.randn(2) * 0.02
			center = random.choice(centers)

			point[0] += center[0]
			point[1] += center[1]

			dataset.append(point)

		dataset = np.array(dataset).astype(np.float32)
		dataset /= 1.414
		yield dataset



#實現可視化的
def generate_image(D, G, xr, epoch):  #xr就是真實的sample出來的x
    """
    Generates and saves a plot of the true distribution, the generator, and the
    critic.
    """
    N_POINTS = 128
    RANGE = 3
    plt.clf()

    points = np.zeros((N_POINTS, N_POINTS, 2), dtype='float32')
    points[:, :, 0] = np.linspace(-RANGE, RANGE, N_POINTS)[:, None]
    points[:, :, 1] = np.linspace(-RANGE, RANGE, N_POINTS)[None, :]
    points = points.reshape((-1, 2))
    # (16384, 2)
    # print('p:', points.shape)

    # draw contour
    with torch.no_grad():
        points = torch.Tensor(points).cuda() # [16384, 2]
        disc_map = D(points).cpu().numpy() # [16384]
    x = y = np.linspace(-RANGE, RANGE, N_POINTS)
    cs = plt.contour(x, y, disc_map.reshape((len(x), len(y))).transpose())
    plt.clabel(cs, inline=1, fontsize=10)
    # plt.colorbar()


    # draw samples
    with torch.no_grad():
        z = torch.randn(batchsz, 2).cuda() # [b, 2]
        samples = G(z).cpu().numpy() # [b, 2]
    plt.scatter(xr[:, 0], xr[:, 1], c='orange', marker='.')
    plt.scatter(samples[:, 0], samples[:, 1], c='green', marker='+')

    viz.matplot(plt, win='contour', opts=dict(title='p(x):%d'%epoch))


def gradient_penalty(D, xr, xf):
    #[b,1]
    t = torch.rand(batchsz, 1).cuda()
    #[b,1]=>[b,2]
    t = t.expand_as(xr)
    #在真實數據和fake數據之間做一個線性插值
    mid = t * xr + (1-t) * xf
    #設置它需要導數信息
    mid.requires_grad_()

    pred = D(mid)
    grads = autograd.grad(outputs=pred, inputs=mid,
                          grad_outputs=torch.ones_like(pred),
                          create_graph=True, retain_graph=True, only_inputs=True)[0]
    gp = torch.pow(grads.norm(2, dim=1)-1,2).mean()

    return gp

def main():
	torch.manual_seed(23)
	np.random.seed(23)

	data_iter = data_generator()
	x = next(data_iter)
	#[b,2]
	# print(x.shape)

	G = Generator().cuda()
	D = Discriminator().cuda()
	# print(G)
	# print(D)

	optim_G = optim.Adam(G.parameters(), lr=5e-4, betas=(0.5, 0.9))
	optim_D = optim.Adam(D.parameters(), lr=5e-4, betas=(0.5, 0.9))

	viz.line([[0,0]], [0], win='loss', opts=dict(title='loss',legend=['D','G']))

	for epoch in range(50000):
		#1、首先訓練Discriminator
		for _ in range(5):
			#1、train on real data
			x = next(data_iter)
			xr = torch.from_numpy(x).cuda()
			#把真實數據送入discriminator
			#[b,2]=>[b,1]
			predr = D(xr)
			#max predr
			lossr = -predr.mean()

			#1.2 train on fake data
			z = torch.randn(batchsz,2).cuda()
			#xf是xfake
			xf = G(z).detach()  #類似於tf.stop_gradient()
			predf = D(xf)
			#min loss
			lossf = predf.mean()


			#1.3 gradient penalty
			gp = gradient_penalty(D, xr, xf.detach())

			#aggregate all
			loss_D = lossr + lossf + 0.2 * gp

			#optimize
			optim_D.zero_grad()
			loss_D.backward()
			optim_D.step()



		#2、訓練Generator
		#從數據中sample一個z出來
		z = torch.randn(batchsz,2).cuda()
		xf = G(z)
		predf = D(xf)
		#max predf.mean()
		loss_G = -predf.mean()

		#optimize
		optim_G.zero_grad()
		loss_G.backward()
		optim_G.step()


		if epoch % 100 == 0:
			viz.line([[loss_D.item(), loss_G.item()]], [epoch], win='loss', update='append')
			print(loss_D.item(), loss_G.item())
			generate_image(D,G, xr.cpu(), epoch)

if __name__=='__main__':
	main()