#!/usr/bin/env python
import os
import numpy as np
from PIL import Image
import chainer
import chainer.cuda
from chainer import Variable
def out_image(updater, enc, dec, rows, cols, seed, dst, gpu, w_img, h_img):
@chainer.training.make_extension()
def make_image(trainer):
np.random.seed(seed)
n_images = rows * cols
xp = enc.xp
in_ch = 3
out_ch = 3
in_all = np.zeros((n_images, in_ch, h_img, w_img)).astype("f")
gt_all = np.zeros((n_images, out_ch, h_img, w_img)).astype("f")
gen_all = np.zeros((n_images, out_ch, h_img, w_img)).astype("f")
for it in range(n_images):
batch = updater.get_iterator('test').next()
batchsize = len(batch)
x_in = xp.zeros((batchsize, in_ch, h_img, w_img)).astype("f")
t_out = xp.zeros((batchsize, out_ch, h_img, w_img)).astype("f")
for i in range(batchsize):
x_in[i,:] = xp.asarray(batch[i][0])
t_out[i,:] = xp.asarray(batch[i][1])
x_in = Variable(x_in)
z = enc(x_in)
x_out = dec(z)
if gpu >= 0:
in_all[it,:] = x_in.data.get()[0,:]
gt_all[it,:] = t_out.get()[0,:]
gen_all[it,:] = x_out.data.get()[0,:]
else:
in_all[it,:] = x_in.data[0,:]
gt_all[it,:] = t_out[0,:]
gen_all[it,:] = x_out.data[0,:]
def save_image(x, name, mode=None):
_, C, H, W = x.shape
x = x.reshape((rows, cols, C, H, W))
x = x.transpose(0, 3, 1, 4, 2)
if C==1:
x = x.reshape((rows*H, cols*W))
else:
x = x.reshape((rows*H, cols*W, C))
preview_dir = '{}/preview'.format(dst)
preview_path = preview_dir +\
'/image_{:0>8}_{}.png'.format(trainer.updater.iteration, name)
if not os.path.exists(preview_dir):
os.makedirs(preview_dir)
Image.fromarray(x, mode=mode).convert('RGB').save(preview_path)
"""
x = np.asarray(np.clip(gen_all, 0.0, 255.0), dtype=np.uint8)
save_image(x, "gen")
x = np.asarray(np.clip(in_all, 0.0, 255.0), dtype=np.uint8)
save_image(x, "in")
x = np.asarray(np.clip(gt_all, 0.0, 255.0), dtype=np.uint8)
save_image(x, "gt")
"""
x = np.asarray(np.clip(gen_all * 128 + 128, 0.0, 255.0), dtype=np.uint8)
save_image(x, "gen")
x = np.asarray(np.clip(in_all * 128+128, 0.0, 255.0), dtype=np.uint8)
save_image(x, "in")
x = np.asarray(np.clip(gt_all * 128+128, 0.0, 255.0), dtype=np.uint8)
save_image(x, "gt")
return make_image