-
Notifications
You must be signed in to change notification settings - Fork 0
/
test.py
148 lines (121 loc) · 4.84 KB
/
test.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
import torch
import torch.nn as nn
import torchvision
import torch.backends.cudnn as cudnn
import torch.optim
import os
import sys
import argparse
import time
import dataloader
import model
import logger, pdb
import numpy as np
import math
run_num = 7
# tb_path = '/media/hdd1/shashant/super_slomo/tb_logs/'
# model_path = os.path.join('/media/hdd1/shashant/super_slomo/baselines/')
model_path = os.path.join('/media/hdd1/shashant/super_slomo/models',str(run_num))
# if not os.path.exists(tb_path):
# os.makedirs(tb_path)
# if not os.listdir(tb_path):
# run_num = 1
# else:
# run_num = max([int(x) for x in os.listdir(tb_path)]) + 1
global_step = 0
class FlowWarper(nn.Module):
def __init__(self, w, h):
super(FlowWarper, self).__init__()
x = np.arange(0,w)
y = np.arange(0,h)
gx, gy = np.meshgrid(x,y)
self.w = w
self.h = h
self.grid_x = torch.autograd.Variable(torch.Tensor(gx), requires_grad=False).cuda()
self.grid_y = torch.autograd.Variable(torch.Tensor(gy), requires_grad=False).cuda()
def forward(self, img, uv):
u = uv[:,0,:,:]
v = uv[:,1,:,:]
X = self.grid_x.unsqueeze(0).expand_as(u) + u # what's happening here?
Y = self.grid_y.unsqueeze(0).expand_as(v) + v
X = 2*(X/self.w - 0.5) # normalize and mean = 0
Y = 2*(Y/self.h - 0.5)
grid_tf = torch.stack((X,Y), dim=3)
img_tf = torch.nn.functional.grid_sample(img, grid_tf) # bilinear interp
return img_tf
def test():
flowModel = model.UNet_flow().cuda()
interpolationModel = model.UNet_refine().cuda()
# Load pretrained flowModel and interpolationModel
flow_model = os.path.join(model_path, 'checkpoint_flow_14_1999.pt')
interp_model = os.path.join(model_path, 'checkpoint_interp_14_1999.pt')
flow_chkpt = torch.load(flow_model)
interp_chkpt = torch.load(interp_model)
flowModel.load_state_dict(flow_chkpt)
interpolationModel.load_state_dict(interp_chkpt)
# ### ResNet for Perceptual Loss
# res50_model = torchvision.models.resnet18(pretrained=True)
# res50_conv = nn.Sequential(*list(res50_model.children())[:-2])
# res50_conv.cuda()
# for param in res50_conv.parameters():
# param.requires_grad = False
dataFeeder = dataloader.testLoader('/media/hdd1/datasets/Adobe240-fps/my_high_fps_frames/slomo_rgb/000055', mode='test')
test_loader = torch.utils.data.DataLoader(dataFeeder, batch_size=1,
shuffle=False, num_workers=1,
pin_memory=True)
flowModel.eval()
interpolationModel.eval()
warper = FlowWarper(352,352)
count = 0
for i, (imageList) in enumerate(test_loader):
I0_var = torch.autograd.Variable(imageList[0]).cuda()
I1_var = torch.autograd.Variable(imageList[-1]).cuda()
flow_out_var = flowModel(I0_var, I1_var)
F_0_1 = flow_out_var[:,:2,:,:]
F_1_0 = flow_out_var[:,2:,:,:]
np.save('./opt_01_0.npy', F_0_1.detach().cpu().numpy())
np.save('./opt_10_0.npy', F_1_0.detach().cpu().numpy())
image_collector = []
for t_ in range(1,8):
t = t_/8
F_t_0 = -(1-t)*t*F_0_1 + t*t*F_1_0
F_t_1 = (1-t)*(1-t)*F_0_1 - t*(1-t)*(F_1_0)
np.save('./opt_t0_'+str(t_)+'.npy', F_t_0.detach().cpu().numpy())
np.save('./opt_t1_'+str(t_)+'.npy', F_t_1.detach().cpu().numpy())
g_I0_F_t_0 = warper(I0_var, F_t_0)
g_I1_F_t_1 = warper(I1_var, F_t_1)
interp_out_var = interpolationModel(I0_var, I1_var, F_0_1, F_1_0, F_t_0, F_t_1, g_I0_F_t_0, g_I1_F_t_1)
F_t_0_final = interp_out_var[:,:2,:,:] + F_t_0
F_t_1_final = interp_out_var[:,2:4,:,:] + F_t_1
V_t_0 = torch.unsqueeze(interp_out_var[:,4,:,:],1)
V_t_1 = 1 - V_t_0
g_I0_F_t_0_final = warper(I0_var, F_t_0_final)
g_I0_F_t_1_final = warper(I1_var, F_t_1_final)
normalization = (1-t)*V_t_0 + t*V_t_1
interpolated_image_t_pre = (1-t)*V_t_0*g_I0_F_t_0_final + t*V_t_1*g_I0_F_t_1_final
interpolated_image_t = interpolated_image_t_pre / normalization
image_collector.append(interpolated_image_t)
save_path = os.path.join('/media/hdd1/datasets/Adobe240-fps/pred_high_fps_frames',str(run_num))
if not os.path.exists(save_path):
os.makedirs(save_path)
torchvision.utils.save_image((I0_var),os.path.join(save_path, str(count).zfill(4) +'.jpg'),normalize=True)
count += 1
for jj,image in enumerate(image_collector):
torchvision.utils.save_image((image),os.path.join(save_path, str(count).zfill(4)+'.jpg'),normalize=True)
count += 1
torchvision.utils.save_image((I1_var),os.path.join(save_path, str(count).zfill(4) +'.jpg'),normalize=True)
count += 1
return save_path
def make_vid(save_path):
frame_path = save_path
vid_path = save_path.replace('frames/', 'videos/')
if not os.path.exists(vid_path):
os.makedirs(vid_path)
os_cmd = 'ffmpeg -framerate 60 -i '+frame_path +'/%04d.jpg '+os.path.join(vid_path)+'/test_sim1.mp4 '
# print(os_cmd)
os.system(os_cmd)
if __name__ == '__main__':
# train_val()
# save_path = os.path.join('/media/hdd1/datasets/Adobe240-fps/pred_high_fps_frames',str(run_num))
save_path = test()
make_vid(save_path)