-
Notifications
You must be signed in to change notification settings - Fork 1
/
logger.py
183 lines (150 loc) · 8.01 KB
/
logger.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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
import numpy as np
import torch
import imageio
import os
from skimage.draw import circle
import matplotlib.pyplot as plt
class Logger:
def __init__(self, log_dir, log_file_name='log.txt', log_freq_iter=100, cpk_freq_epoch=100,
zfill_num=8, visualizer_params=None):
self.loss_list = []
self.cpk_dir = log_dir
self.visualizations_dir = os.path.join(log_dir, 'train-vis')
if not os.path.exists(self.visualizations_dir):
os.makedirs(self.visualizations_dir)
self.log_file = open(os.path.join(log_dir, log_file_name), 'a')
self.log_freq = log_freq_iter
self.cpk_freq = cpk_freq_epoch
self.zfill_num = zfill_num
self.visualizer = Visualizer(**visualizer_params)
self.epoch = 0
self.it = 0
def log_scores(self, loss_names):
loss_mean = np.array(self.loss_list).mean(axis=0)
print(loss_names)
print(loss_mean)
loss_string = "; ".join(["%s - %.5f" % (name, value) for name, value in zip(loss_names, loss_mean)])
loss_string = str(self.it).zfill(self.zfill_num) + ") " + loss_string
print(loss_string, file=self.log_file)
self.loss_list = []
self.log_file.flush()
def visualize_rec(self, inp, out):
image = self.visualizer.visualize_reconstruction(inp, out)
imageio.mimsave(os.path.join(self.visualizations_dir, "%s-rec.gif" % str(self.it).zfill(self.zfill_num)), image)
def save_cpk(self):
cpk = {k: v.state_dict() for k, v in self.models.items()}
cpk['epoch'] = self.epoch
cpk['it'] = self.it
torch.save(cpk, os.path.join(self.cpk_dir, '%s-checkpoint.pth.tar' % str(self.epoch).zfill(self.zfill_num)))
@staticmethod
def load_cpk(checkpoint_path, generator=None, discriminator=None, kp_detector=None,
optimizer_generator=None, optimizer_discriminator=None, optimizer_kp_detector=None):
if torch.cuda.is_available():
checkpoint = torch.load(checkpoint_path)
else:
checkpoint = torch.load(checkpoint_path, map_location='cpu')
if generator is not None:
generator.load_state_dict(checkpoint['generator'])
if kp_detector is not None:
kp_detector.load_state_dict(checkpoint['kp_detector'])
if discriminator is not None:
discriminator.load_state_dict(checkpoint['discriminator'])
if optimizer_generator is not None:
optimizer_generator.load_state_dict(checkpoint['optimizer_generator'])
if optimizer_discriminator is not None:
optimizer_discriminator.load_state_dict(checkpoint['optimizer_discriminator'])
if optimizer_kp_detector is not None:
optimizer_kp_detector.load_state_dict(checkpoint['optimizer_kp_detector'])
return checkpoint['epoch'], checkpoint['it']
def __enter__(self):
return self
def __exit__(self, exc_type, exc_val, exc_tb):
if 'models' in self.__dict__:
self.save_cpk()
self.log_file.close()
def log_iter(self, it, names, values, inp, out = None):
self.it = it
self.names = names
self.loss_list.append(values)
if it % self.log_freq == 0:
self.log_scores(self.names)
if out is not None:
self.visualize_rec(inp, out)
def log_epoch(self, epoch, models):
self.epoch = epoch
self.models = models
if epoch % self.cpk_freq == 0:
self.save_cpk()
class Visualizer:
def __init__(self, kp_size=2, draw_border=False, colormap='gist_rainbow'):
self.kp_size = kp_size
self.draw_border = draw_border
self.colormap = plt.get_cmap(colormap)
def draw_video_with_kp(self, video, kp_array):
video_array = np.copy(video)
spatial_size = np.array(video_array.shape[2:0:-1])[np.newaxis, np.newaxis]
kp_array = spatial_size * (kp_array + 1) / 2
num_kp = kp_array.shape[1]
for i in range(len(video_array)):
for kp_ind, kp in enumerate(kp_array[i]):
rr, cc = circle(kp[1], kp[0], self.kp_size, shape=video_array.shape[1:3])
video_array[i][rr, cc] = np.array(self.colormap(kp_ind / num_kp))[:3]
return video_array
def create_video_column_with_kp(self, video, kp):
video_array = np.array([self.draw_video_with_kp(v, k) for v, k in zip(video, kp)])
return self.create_video_column(video_array)
def create_video_column(self, videos):
if self.draw_border:
videos = np.copy(videos)
videos[:, :, [0, -1]] = (1, 1, 1)
videos[:, :, :, [0, -1]] = (1, 1, 1)
return np.concatenate(list(videos), axis=1)
def create_image_grid(self, *args):
out = []
for arg in args:
if type(arg) == tuple:
out.append(self.create_video_column_with_kp(arg[0], arg[1]))
else:
out.append(self.create_video_column(arg))
return np.concatenate(out, axis=2)
def visualize_transfer(self, driving_video, source_image, out):
out_video_batch = out['video_prediction'].data.cpu().numpy()
appearance_deformed_batch = out['video_deformed'].data.cpu().numpy()
motion_video_batch = driving_video.data.cpu().numpy()
appearance_video_batch = source_image[:, :, 0:1].data.cpu().repeat(1, 1, out_video_batch.shape[2],
1, 1).numpy()
video_first_frame = driving_video[:, :, 0:1].data.cpu().repeat(1, 1, out_video_batch.shape[2], 1,
1).numpy()
kp_video = out['kp_driving']['mean'].data.cpu().numpy()
kp_appearance = out['kp_source']['mean'].data.cpu().repeat(1, out_video_batch.shape[2], 1, 1).numpy()
kp_norm = out['kp_norm']['mean'].data.cpu().numpy()
kp_video_first = out['kp_driving']['mean'][:, :1].data.cpu().repeat(1, out_video_batch.shape[2], 1, 1).numpy()
video_first_frame = np.transpose(video_first_frame, [0, 2, 3, 4, 1])
out_video_batch = np.transpose(out_video_batch, [0, 2, 3, 4, 1])
motion_video_batch = np.transpose(motion_video_batch, [0, 2, 3, 4, 1])
appearance_video_batch = np.transpose(appearance_video_batch, [0, 2, 3, 4, 1])
appearance_deformed_batch = np.transpose(appearance_deformed_batch, [0, 2, 3, 4, 1])
image = self.create_image_grid((appearance_video_batch, kp_appearance), (video_first_frame, kp_video_first),
(motion_video_batch, kp_video),
(out_video_batch, kp_norm), out_video_batch, appearance_deformed_batch)
image = (255 * image).astype(np.uint8)
return image
def visualize_reconstruction(self, inp, out):
out_video_batch = out['video_prediction'].data.cpu().numpy()
if 'driving' in inp:
gt_video_batch = inp['driving'].data.cpu().numpy()
else:
gt_video_batch = inp['video'].data.cpu().numpy()
appearance_deformed_batch = out['video_deformed'].data.cpu().numpy()
appearance_video_batch = inp['source'].data.cpu().repeat(1, 1, out_video_batch.shape[2], 1, 1).numpy()
kp_video = out['kp_driving']['mean'].data.cpu().numpy()
kp_appearance = out['kp_source']['mean'].data.cpu().repeat(1, out_video_batch.shape[2], 1, 1).numpy()
out_video_batch = np.transpose(out_video_batch, [0, 2, 3, 4, 1])
gt_video_batch = np.transpose(gt_video_batch, [0, 2, 3, 4, 1])
appearance_video_batch = np.transpose(appearance_video_batch, [0, 2, 3, 4, 1])
appearance_deformed_batch = np.transpose(appearance_deformed_batch, [0, 2, 3, 4, 1])
image = self.create_image_grid((appearance_video_batch, kp_appearance), (gt_video_batch, kp_video),
out_video_batch,
appearance_deformed_batch, gt_video_batch)
image = (255 * image).astype(np.uint8)
return image