-
Notifications
You must be signed in to change notification settings - Fork 2
/
inference_acgan.py
47 lines (41 loc) · 1.85 KB
/
inference_acgan.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
import sys
import os
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
session = tf.Session(config=config)
import keras
from keras.utils import to_categorical
from keras import backend as K
K.set_session(session)
from keras.models import load_model
from models import up_bilinear
from skimage.io import imsave
from pixel_shuffler import PixelShuffler
import pandas as pd
import argparse
parser = argparse.ArgumentParser(description='Image Generation with GAN')
parser.add_argument('output', metavar='output', type=str, help='')
parser.add_argument('label', metavar='label', type=str, help='')
parser.add_argument('tag_file', metavar='tag_file', type=str, help='')
parser.add_argument('--model', type=str, default='./decoder.h5', required=False, help='model')
parser.add_argument('--n', type=int, default=64, required=False, help='')
parser.add_argument('--std', type=float, default=0.7, required=False, help='')
parser.add_argument('--batch_size', type=int, default=8, required=False, help='')
args = parser.parse_args()
tags = pd.read_csv(args.tag_file) # order is manner
mask = (tags['tags']==args.label)
assert mask.any(), '%s is not in \'%s\'!'%(args.label, args.tag_file)
label_idx = np.argmax(mask)
if not os.path.exists(args.output):
os.makedirs(args.output)
model = load_model(args.model, custom_objects={'tf':tf, 'PixelShuffler':PixelShuffler, 'up_bilinear':up_bilinear})
N_CLASS = len(tags['tags'])
noise = np.random.normal(0, args.std, (args.n, model.input_shape[-1]-N_CLASS))
labels = to_categorical(np.asarray([label_idx]*args.n), N_CLASS)
z = np.append(noise, labels, axis=-1)
m = (model.predict(z, batch_size=args.batch_size) * 127.5 + 127.5).astype(np.uint8)
for i in range(args.n):
imsave(os.path.join(args.output, 'output_{:04d}.png'.format(i)), np.squeeze(m[i]))