-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathtest.py
executable file
·178 lines (146 loc) · 6.21 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
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
import argparse
import os
import math
from functools import partial
import yaml
import torch
from torch.utils.data import DataLoader
from tqdm import tqdm
import torch.nn as nn
import datasets
import models
import utils
from skimage.metrics import peak_signal_noise_ratio as PSNR
from skimage.metrics import structural_similarity as SSIM
from torchvision.utils import save_image
from utils import createDirectory,foldername,filename
def make_data_loader(spec, tag=''):
if spec is None:
return None
dataset = datasets.make(spec['dataset'])
dataset = datasets.make(spec['wrapper'], args={'dataset': dataset})
log('{} dataset: size={}'.format(tag, len(dataset)))
for k, v in dataset[0].items():
log(' {}: shape={}'.format(k, tuple(v.shape)))
loader = DataLoader(dataset, batch_size=spec['batch_size'],
shuffle=(tag == 'train'), num_workers=8, pin_memory=True)
return loader
def make_data_loaders():
val_loader = make_data_loader(config.get('val_dataset'), tag='val')
return val_loader
def batched_predict(model, inp, coord, cell, bsize):#30000
with torch.no_grad():
model.gen_feat(inp)
n = coord.shape[1]#bs*q
ql = 0
preds = []
while ql < n:
qr = min(ql + bsize, n)
pred = model.query_rgb(coord[:, ql: qr, :], cell[:, ql: qr, :])
preds.append(pred)
ql = qr
pred = torch.cat(preds, dim=1)
return pred
def eval_psnr(loader, model, eval_type=None, eval_bsize=None,
window_size=0, lowbit = 4, highbit = 16, save= 0, save_path = '',verbose=False):
model.eval()
metric_fn = utils.calc_psnr
val_res = utils.Averager()
val_ssim = utils.Averager()
basis = 2**(highbit-lowbit)/((2**highbit) -1)
gtdepth =(2**highbit) -1
pbar = tqdm(loader, leave=False, desc='val')
i=0
with torch.no_grad():
for batch in pbar:
for k, v in batch.items():
batch[k] = v.cuda()
inp = (batch['inp'])
if window_size is not 0:
_, _, h_old, w_old = inp.size()
h_pad = (h_old // window_size + 1) * window_size - h_old
w_pad = (w_old // window_size + 1) * window_size - w_old
inp = torch.cat([inp, torch.flip(inp, [2])], 2)[:, :, :h_old + h_pad, :]
inp = torch.cat([inp, torch.flip(inp, [3])], 3)[:, :, :, :w_old + w_pad]
coord = utils.make_coord((h_old+h_pad,w_old+w_pad)).unsqueeze(0).cuda()
cell = torch.ones_like(coord[:,:,0]).unsqueeze(-1).cuda()
cell = cell*basis
else:
h_pad = 0
w_pad = 0
coord = batch['coord']
cell = batch['cell']
gt = batch['gt']
h,w = inp.shape[-2:]
pred = model(inp, coord, cell).view(-1,h, w, 3).permute(0,3, 1, 2)
hdimage = pred*basis+inp
hdimage = hdimage.clamp(0,1)
if window_size is not 0:
gt = gt.view(-1,h_old,w_old,3).permute(0,3,1,2)
inp = inp[:,:,:h_old,:w_old]
hdimage=hdimage[:,:,:h_old,:w_old]
else:
gt = gt.view(-1,h,w,3).permute(0,3,1,2)
gt_ = gt*basis+inp
val_ssim.add(SSIM(hdimage.squeeze().permute(1,2,0).cpu().numpy()*gtdepth,\
gt_.squeeze().permute(1,2,0).cpu().numpy()*gtdepth,\
channel_axis=0,data_range=gtdepth,multichannel=True),\
inp.shape[0])
val_res.add(PSNR(hdimage.squeeze().cpu().numpy(),gt_.squeeze().cpu().numpy()),inp.shape[0])
if save is 1:
inp = inp.squeeze()
hdimage = hdimage.squeeze()
gt_ = gt_.squeeze()
save_image(hdimage ,save_path +'/'+str(i)+'_'+str(lowbit)+'_'+str(highbit)+'_BDE_ABCD'+'.png')
save_image(gt_ ,save_path +'/'+str(i)+'_'+str(highbit)+'_GT'+'.png')
save_image(inp ,save_path +'/'+str(i)+'_'+str(lowbit)+'_input'+'.png')
i = i+1
if verbose:
pbar.set_description('PSNR {:.4f}'.format(val_res.item()),'SSIM {:.4f}'.format(val_ssim.item()))
return val_res.item(), val_ssim.item()
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--config')
parser.add_argument('--model')
parser.add_argument('--window', default='0')
parser.add_argument('--LBD', default='4')
parser.add_argument('--HBD', default='16')
parser.add_argument('--save', default='0')
parser.add_argument('--foldertag', default='')
parser.add_argument('--gpu', default='0')
args = parser.parse_args()
os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
q =int(args.LBD)
N =int(args.HBD)
windowsize = int(args.window )
with open(args.config, 'r') as f:
config = yaml.load(f, Loader=yaml.FullLoader)
spec = config['test_dataset']
dataset = datasets.make(spec['dataset'])
dataset = datasets.make(spec['wrapper'], args={'dataset': dataset,'inpdepth': q,'gtdepth':N})
loader = DataLoader(dataset, batch_size=spec['batch_size'],
num_workers=8, pin_memory=True)
#make_data_loader(config.get('val_dataset'), tag='val')
model_spec = torch.load(args.model)['model']
model = models.make(model_spec, load_sd=True).cuda()
n_gpus = len(os.environ['CUDA_VISIBLE_DEVICES'].split(','))
if n_gpus > 1:
model = nn.parallel.DataParallel(model)
save_ = int(args.save)
tag = args.foldertag
save_path = ''
if save_ is 1:
folder_name = foldername(args.config)##ex)kodak
model_name = foldername(args.model)##ex)train_~~
save_path='./result/'+model_name+'/'+str(q)+'-'+str(N)+'/'+folder_name + tag
createDirectory(save_path)
res,resssim = eval_psnr(loader, model,
eval_type=config.get('eval_type'),
eval_bsize=config.get('eval_bsize'),
lowbit = q,
highbit = N,
window_size= windowsize,
save=save_,
save_path = save_path,
verbose=True)
print('PSNR: {:.4f}'.format(res),'SSIM: {:.4f}'.format(resssim))