OAOA commited on
Commit
bfa59ab
1 Parent(s): 25c14b8

first commit

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. LICENSE +35 -0
  2. README.md +3 -0
  3. app.py +141 -0
  4. basicsr/.DS_Store +0 -0
  5. basicsr/__init__.py +4 -0
  6. basicsr/data/.DS_Store +0 -0
  7. basicsr/data/__init__.py +101 -0
  8. basicsr/data/__pycache__/__init__.cpython-310.pyc +0 -0
  9. basicsr/data/__pycache__/__init__.cpython-38.pyc +0 -0
  10. basicsr/data/__pycache__/data_util.cpython-310.pyc +0 -0
  11. basicsr/data/__pycache__/data_util.cpython-38.pyc +0 -0
  12. basicsr/data/__pycache__/degradations.cpython-310.pyc +0 -0
  13. basicsr/data/__pycache__/degradations.cpython-38.pyc +0 -0
  14. basicsr/data/__pycache__/ffhq_dataset.cpython-310.pyc +0 -0
  15. basicsr/data/__pycache__/ffhq_dataset.cpython-38.pyc +0 -0
  16. basicsr/data/__pycache__/paired_image_dataset.cpython-310.pyc +0 -0
  17. basicsr/data/__pycache__/paired_image_dataset.cpython-38.pyc +0 -0
  18. basicsr/data/__pycache__/prefetch_dataloader.cpython-310.pyc +0 -0
  19. basicsr/data/__pycache__/prefetch_dataloader.cpython-38.pyc +0 -0
  20. basicsr/data/__pycache__/realesrgan_dataset.cpython-310.pyc +0 -0
  21. basicsr/data/__pycache__/realesrgan_dataset.cpython-38.pyc +0 -0
  22. basicsr/data/__pycache__/realesrgan_paired_dataset.cpython-310.pyc +0 -0
  23. basicsr/data/__pycache__/realesrgan_paired_dataset.cpython-38.pyc +0 -0
  24. basicsr/data/__pycache__/reds_dataset.cpython-310.pyc +0 -0
  25. basicsr/data/__pycache__/reds_dataset.cpython-38.pyc +0 -0
  26. basicsr/data/__pycache__/single_image_dataset.cpython-310.pyc +0 -0
  27. basicsr/data/__pycache__/single_image_dataset.cpython-38.pyc +0 -0
  28. basicsr/data/__pycache__/transforms.cpython-310.pyc +0 -0
  29. basicsr/data/__pycache__/transforms.cpython-38.pyc +0 -0
  30. basicsr/data/__pycache__/video_test_dataset.cpython-310.pyc +0 -0
  31. basicsr/data/__pycache__/video_test_dataset.cpython-38.pyc +0 -0
  32. basicsr/data/__pycache__/vimeo90k_dataset.cpython-310.pyc +0 -0
  33. basicsr/data/__pycache__/vimeo90k_dataset.cpython-38.pyc +0 -0
  34. basicsr/data/data_sampler.py +48 -0
  35. basicsr/data/data_util.py +315 -0
  36. basicsr/data/degradations.py +765 -0
  37. basicsr/data/ffhq_dataset.py +80 -0
  38. basicsr/data/meta_info/meta_info_DIV2K800sub_GT.txt +0 -0
  39. basicsr/data/meta_info/meta_info_REDS4_test_GT.txt +4 -0
  40. basicsr/data/meta_info/meta_info_REDS_GT.txt +270 -0
  41. basicsr/data/meta_info/meta_info_REDSofficial4_test_GT.txt +4 -0
  42. basicsr/data/meta_info/meta_info_REDSval_official_test_GT.txt +30 -0
  43. basicsr/data/meta_info/meta_info_Vimeo90K_test_GT.txt +0 -0
  44. basicsr/data/meta_info/meta_info_Vimeo90K_test_fast_GT.txt +1225 -0
  45. basicsr/data/meta_info/meta_info_Vimeo90K_test_medium_GT.txt +0 -0
  46. basicsr/data/meta_info/meta_info_Vimeo90K_test_slow_GT.txt +1613 -0
  47. basicsr/data/meta_info/meta_info_Vimeo90K_train_GT.txt +0 -0
  48. basicsr/data/paired_image_dataset.py +106 -0
  49. basicsr/data/prefetch_dataloader.py +122 -0
  50. basicsr/data/realesrgan_dataset.py +384 -0
LICENSE ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ S-Lab License 1.0
2
+
3
+ Copyright 2024 S-Lab
4
+
5
+ Redistribution and use for non-commercial purpose in source and
6
+ binary forms, with or without modification, are permitted provided
7
+ that the following conditions are met:
8
+
9
+ 1. Redistributions of source code must retain the above copyright
10
+ notice, this list of conditions and the following disclaimer.
11
+
12
+ 2. Redistributions in binary form must reproduce the above copyright
13
+ notice, this list of conditions and the following disclaimer in
14
+ the documentation and/or other materials provided with the
15
+ distribution.
16
+
17
+ 3. Neither the name of the copyright holder nor the names of its
18
+ contributors may be used to endorse or promote products derived
19
+ from this software without specific prior written permission.
20
+
21
+ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22
+ "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23
+ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
24
+ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
25
+ HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
26
+ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
27
+ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
28
+ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
29
+ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
30
+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
31
+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32
+
33
+ In the event that redistribution and/or use for commercial purpose in
34
+ source or binary forms, with or without modification is required,
35
+ please contact the contributor(s) of the work.
README.md CHANGED
@@ -5,6 +5,9 @@ colorFrom: green
5
  colorTo: purple
6
  sdk: gradio
7
  sdk_version: 5.8.0
 
 
 
8
  app_file: app.py
9
  pinned: false
10
  license: other
 
5
  colorTo: purple
6
  sdk: gradio
7
  sdk_version: 5.8.0
8
+ python_version: 3.10
9
+ suggested_storage: small
10
+ models: OAOA/InvSR
11
  app_file: app.py
12
  pinned: false
13
  license: other
app.py ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python
2
+ # -*- coding:utf-8 -*-
3
+ # Power by Zongsheng Yue 2024-12-11 17:17:41
4
+
5
+ import spaces
6
+ import warnings
7
+ warnings.filterwarnings("ignore")
8
+
9
+ import argparse
10
+ import numpy as np
11
+ import gradio as gr
12
+ from pathlib import Path
13
+ from omegaconf import OmegaConf
14
+ from sampler_invsr import InvSamplerSR
15
+
16
+ from utils import util_common
17
+ from utils import util_image
18
+ from basicsr.utils.download_util import load_file_from_url
19
+
20
+ def get_configs(num_steps=1, chopping_size=128, seed=12345):
21
+ configs = OmegaConf.load("./configs/sample-sd-turbo.yaml")
22
+
23
+ if num_steps == 1:
24
+ configs.timesteps = [200,]
25
+ elif num_steps == 2:
26
+ configs.timesteps = [200, 100]
27
+ elif num_steps == 3:
28
+ configs.timesteps = [200, 100, 50]
29
+ elif num_steps == 4:
30
+ configs.timesteps = [200, 150, 100, 50]
31
+ elif num_steps == 5:
32
+ configs.timesteps = [250, 200, 150, 100, 50]
33
+ else:
34
+ assert num_steps <= 250
35
+ configs.timesteps = np.linspace(
36
+ start=250, stop=0, num=num_steps, endpoint=False, dtype=np.int64()
37
+ ).tolist()
38
+ print(f'Setting timesteps for inference: {configs.timesteps}')
39
+
40
+ # path to save noise predictor
41
+ started_ckpt_path = "noise_predictor_sd_turbo_v5.pth"
42
+ # started_ckpt_dir = "./weights"
43
+ # util_common.mkdir(started_ckpt_dir, delete=False, parents=True)
44
+ # started_ckpt_path = Path(started_ckpt_dir) / started_ckpt_name
45
+ # if not started_ckpt_path.exists():
46
+ # load_file_from_url(
47
+ # url="https://huggingface.co/OAOA/InvSR/resolve/main/noise_predictor_sd_turbo_v5.pth",
48
+ # model_dir=started_ckpt_dir,
49
+ # progress=True,
50
+ # file_name=started_ckpt_name,
51
+ # )
52
+ configs.model_start.ckpt_path = started_ckpt_path
53
+
54
+ configs.bs = 1
55
+ configs.seed = 12345
56
+ configs.basesr.chopping.pch_size = chopping_size
57
+
58
+ return configs
59
+
60
+ @spaces.GPU
61
+ def predict(in_path, num_steps=1, chopping_size=128, seed=12345):
62
+ configs = get_configs(num_steps=num_steps, chopping_size=chopping_size, seed=12345)
63
+
64
+ sampler = InvSamplerSR(configs)
65
+
66
+ out_dir = Path('invsr_output')
67
+ if not out_dir.exists():
68
+ out_dir.mkdir()
69
+ sampler.inference(in_path, out_path=out_dir, bs=1)
70
+
71
+ out_path = out_dir / f"{Path(in_path).stem}.png"
72
+ assert out_path.exists(), 'Super-resolution failed!'
73
+ im_sr = util_image.imread(out_path, chn="rgb", dtype="uint8")
74
+
75
+ return im_sr, str(out_path)
76
+
77
+ title = "Arbitrary-steps Image Super-resolution via Diffusion Inversion"
78
+ description = r"""
79
+ <b>Official Gradio demo</b> for <a href='https://github.com/zsyOAOA/InvSR' target='_blank'><b>Arbitrary-steps Image Super-resolution via Diffuion Inversion</b></a>.<br>
80
+ 🔥 InvSR is an image super-resolution method via Diffusion Inversion, supporting arbitrary sampling steps.<br>
81
+ """
82
+ article = r"""
83
+ If you've found InvSR useful for your research or projects, please show your support by ⭐ the <a href='https://github.com/zsyOAOA/InvSR' target='_blank'>Github Repo</a>. Thanks!
84
+ [![GitHub Stars](https://img.shields.io/github/stars/zsyOAOA/InvSR?affiliations=OWNER&color=green&style=social)](https://github.com/zsyOAOA/InvSR)
85
+
86
+ ---
87
+ If our work is useful for your research, please consider citing:
88
+ ```bibtex
89
+ @article{yue2024InvSR,
90
+ title={Arbitrary-steps Image Super-resolution via Diffusion Inversion},
91
+ author={Yue, Zongsheng and Kang, Liao and Loy, Chen Change},
92
+ journal = {arXiv preprint arXiv:2412.09013},
93
+ year={2024},
94
+ }
95
+ ```
96
+
97
+ 📋 **License**
98
+
99
+ This project is licensed under <a rel="license" href="https://github.com/zsyOAOA/InvSR/blob/master/LICENSE">S-Lab License 1.0</a>.
100
+ Redistribution and use for non-commercial purposes should follow this license.
101
+
102
+ 📧 **Contact**
103
+
104
+ If you have any questions, please feel free to contact me via <b>[email protected]</b>.
105
+ ![visitors](https://visitor-badge.laobi.icu/badge?page_id=zsyOAOA/InvSR)
106
+ """
107
+ demo = gr.Interface(
108
+ fn=predict,
109
+ inputs=[
110
+ gr.Image(type="filepath", label="Input: Low Quality Image"),
111
+ gr.Dropdown(
112
+ choices=[1,2,3,4,5],
113
+ value=1,
114
+ label="Number of steps",
115
+ ),
116
+ gr.Dropdown(
117
+ choices=[128, 256],
118
+ value=128,
119
+ label="Chopping size",
120
+ ),
121
+ gr.Number(value=12345, precision=0, label="Ranom seed")
122
+ ],
123
+ outputs=[
124
+ gr.Image(type="numpy", label="Output: High Quality Image"),
125
+ gr.File(label="Download the output")
126
+ ],
127
+ title=title,
128
+ description=description,
129
+ article=article,
130
+ examples=[
131
+ ['./testdata/RealSet80/29.jpg', 3, 128, 12345],
132
+ ['./testdata/RealSet80/32.jpg', 1, 128, 12345],
133
+ ['./testdata/RealSet80/0030.jpg', 1, 128, 12345],
134
+ ['./testdata/RealSet80/2684538-PH.jpg', 1, 128, 12345],
135
+ ['./testdata/RealSet80/oldphoto6.png', 1, 128, 12345],
136
+ ]
137
+ )
138
+
139
+ demo.queue(max_size=5)
140
+ demo.launch(share=True)
141
+
basicsr/.DS_Store ADDED
Binary file (6.15 kB). View file
 
basicsr/__init__.py ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ # https://github.com/xinntao/BasicSR
2
+ # flake8: noqa
3
+ from .data import *
4
+ from .utils import *
basicsr/data/.DS_Store ADDED
Binary file (6.15 kB). View file
 
basicsr/data/__init__.py ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import importlib
2
+ import numpy as np
3
+ import random
4
+ import torch
5
+ import torch.utils.data
6
+ from copy import deepcopy
7
+ from functools import partial
8
+ from os import path as osp
9
+
10
+ from basicsr.data.prefetch_dataloader import PrefetchDataLoader
11
+ from basicsr.utils import get_root_logger, scandir
12
+ from basicsr.utils.dist_util import get_dist_info
13
+ from basicsr.utils.registry import DATASET_REGISTRY
14
+
15
+ __all__ = ['build_dataset', 'build_dataloader']
16
+
17
+ # automatically scan and import dataset modules for registry
18
+ # scan all the files under the data folder with '_dataset' in file names
19
+ data_folder = osp.dirname(osp.abspath(__file__))
20
+ dataset_filenames = [osp.splitext(osp.basename(v))[0] for v in scandir(data_folder) if v.endswith('_dataset.py')]
21
+ # import all the dataset modules
22
+ _dataset_modules = [importlib.import_module(f'basicsr.data.{file_name}') for file_name in dataset_filenames]
23
+
24
+
25
+ def build_dataset(dataset_opt):
26
+ """Build dataset from options.
27
+
28
+ Args:
29
+ dataset_opt (dict): Configuration for dataset. It must contain:
30
+ name (str): Dataset name.
31
+ type (str): Dataset type.
32
+ """
33
+ dataset_opt = deepcopy(dataset_opt)
34
+ dataset = DATASET_REGISTRY.get(dataset_opt['type'])(dataset_opt)
35
+ logger = get_root_logger()
36
+ logger.info(f'Dataset [{dataset.__class__.__name__}] - {dataset_opt["name"]} is built.')
37
+ return dataset
38
+
39
+
40
+ def build_dataloader(dataset, dataset_opt, num_gpu=1, dist=False, sampler=None, seed=None):
41
+ """Build dataloader.
42
+
43
+ Args:
44
+ dataset (torch.utils.data.Dataset): Dataset.
45
+ dataset_opt (dict): Dataset options. It contains the following keys:
46
+ phase (str): 'train' or 'val'.
47
+ num_worker_per_gpu (int): Number of workers for each GPU.
48
+ batch_size_per_gpu (int): Training batch size for each GPU.
49
+ num_gpu (int): Number of GPUs. Used only in the train phase.
50
+ Default: 1.
51
+ dist (bool): Whether in distributed training. Used only in the train
52
+ phase. Default: False.
53
+ sampler (torch.utils.data.sampler): Data sampler. Default: None.
54
+ seed (int | None): Seed. Default: None
55
+ """
56
+ phase = dataset_opt['phase']
57
+ rank, _ = get_dist_info()
58
+ if phase == 'train':
59
+ if dist: # distributed training
60
+ batch_size = dataset_opt['batch_size_per_gpu']
61
+ num_workers = dataset_opt['num_worker_per_gpu']
62
+ else: # non-distributed training
63
+ multiplier = 1 if num_gpu == 0 else num_gpu
64
+ batch_size = dataset_opt['batch_size_per_gpu'] * multiplier
65
+ num_workers = dataset_opt['num_worker_per_gpu'] * multiplier
66
+ dataloader_args = dict(
67
+ dataset=dataset,
68
+ batch_size=batch_size,
69
+ shuffle=False,
70
+ num_workers=num_workers,
71
+ sampler=sampler,
72
+ drop_last=True)
73
+ if sampler is None:
74
+ dataloader_args['shuffle'] = True
75
+ dataloader_args['worker_init_fn'] = partial(
76
+ worker_init_fn, num_workers=num_workers, rank=rank, seed=seed) if seed is not None else None
77
+ elif phase in ['val', 'test']: # validation
78
+ dataloader_args = dict(dataset=dataset, batch_size=1, shuffle=False, num_workers=0)
79
+ else:
80
+ raise ValueError(f"Wrong dataset phase: {phase}. Supported ones are 'train', 'val' and 'test'.")
81
+
82
+ dataloader_args['pin_memory'] = dataset_opt.get('pin_memory', False)
83
+ dataloader_args['persistent_workers'] = dataset_opt.get('persistent_workers', False)
84
+
85
+ prefetch_mode = dataset_opt.get('prefetch_mode')
86
+ if prefetch_mode == 'cpu': # CPUPrefetcher
87
+ num_prefetch_queue = dataset_opt.get('num_prefetch_queue', 1)
88
+ logger = get_root_logger()
89
+ logger.info(f'Use {prefetch_mode} prefetch dataloader: num_prefetch_queue = {num_prefetch_queue}')
90
+ return PrefetchDataLoader(num_prefetch_queue=num_prefetch_queue, **dataloader_args)
91
+ else:
92
+ # prefetch_mode=None: Normal dataloader
93
+ # prefetch_mode='cuda': dataloader for CUDAPrefetcher
94
+ return torch.utils.data.DataLoader(**dataloader_args)
95
+
96
+
97
+ def worker_init_fn(worker_id, num_workers, rank, seed):
98
+ # Set the worker seed to num_workers * rank + worker_id + seed
99
+ worker_seed = num_workers * rank + worker_id + seed
100
+ np.random.seed(worker_seed)
101
+ random.seed(worker_seed)
basicsr/data/__pycache__/__init__.cpython-310.pyc ADDED
Binary file (3.58 kB). View file
 
basicsr/data/__pycache__/__init__.cpython-38.pyc ADDED
Binary file (3.59 kB). View file
 
basicsr/data/__pycache__/data_util.cpython-310.pyc ADDED
Binary file (11.2 kB). View file
 
basicsr/data/__pycache__/data_util.cpython-38.pyc ADDED
Binary file (11.2 kB). View file
 
basicsr/data/__pycache__/degradations.cpython-310.pyc ADDED
Binary file (20.3 kB). View file
 
basicsr/data/__pycache__/degradations.cpython-38.pyc ADDED
Binary file (21.7 kB). View file
 
basicsr/data/__pycache__/ffhq_dataset.cpython-310.pyc ADDED
Binary file (3.08 kB). View file
 
basicsr/data/__pycache__/ffhq_dataset.cpython-38.pyc ADDED
Binary file (3.02 kB). View file
 
basicsr/data/__pycache__/paired_image_dataset.cpython-310.pyc ADDED
Binary file (3.88 kB). View file
 
basicsr/data/__pycache__/paired_image_dataset.cpython-38.pyc ADDED
Binary file (3.88 kB). View file
 
basicsr/data/__pycache__/prefetch_dataloader.cpython-310.pyc ADDED
Binary file (4.36 kB). View file
 
basicsr/data/__pycache__/prefetch_dataloader.cpython-38.pyc ADDED
Binary file (4.37 kB). View file
 
basicsr/data/__pycache__/realesrgan_dataset.cpython-310.pyc ADDED
Binary file (9.39 kB). View file
 
basicsr/data/__pycache__/realesrgan_dataset.cpython-38.pyc ADDED
Binary file (5.1 kB). View file
 
basicsr/data/__pycache__/realesrgan_paired_dataset.cpython-310.pyc ADDED
Binary file (4.06 kB). View file
 
basicsr/data/__pycache__/realesrgan_paired_dataset.cpython-38.pyc ADDED
Binary file (4.03 kB). View file
 
basicsr/data/__pycache__/reds_dataset.cpython-310.pyc ADDED
Binary file (10.6 kB). View file
 
basicsr/data/__pycache__/reds_dataset.cpython-38.pyc ADDED
Binary file (10.7 kB). View file
 
basicsr/data/__pycache__/single_image_dataset.cpython-310.pyc ADDED
Binary file (2.85 kB). View file
 
basicsr/data/__pycache__/single_image_dataset.cpython-38.pyc ADDED
Binary file (2.82 kB). View file
 
basicsr/data/__pycache__/transforms.cpython-310.pyc ADDED
Binary file (6 kB). View file
 
basicsr/data/__pycache__/transforms.cpython-38.pyc ADDED
Binary file (6.04 kB). View file
 
basicsr/data/__pycache__/video_test_dataset.cpython-310.pyc ADDED
Binary file (10.1 kB). View file
 
basicsr/data/__pycache__/video_test_dataset.cpython-38.pyc ADDED
Binary file (10.3 kB). View file
 
basicsr/data/__pycache__/vimeo90k_dataset.cpython-310.pyc ADDED
Binary file (5.8 kB). View file
 
basicsr/data/__pycache__/vimeo90k_dataset.cpython-38.pyc ADDED
Binary file (5.79 kB). View file
 
basicsr/data/data_sampler.py ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import math
2
+ import torch
3
+ from torch.utils.data.sampler import Sampler
4
+
5
+
6
+ class EnlargedSampler(Sampler):
7
+ """Sampler that restricts data loading to a subset of the dataset.
8
+
9
+ Modified from torch.utils.data.distributed.DistributedSampler
10
+ Support enlarging the dataset for iteration-based training, for saving
11
+ time when restart the dataloader after each epoch
12
+
13
+ Args:
14
+ dataset (torch.utils.data.Dataset): Dataset used for sampling.
15
+ num_replicas (int | None): Number of processes participating in
16
+ the training. It is usually the world_size.
17
+ rank (int | None): Rank of the current process within num_replicas.
18
+ ratio (int): Enlarging ratio. Default: 1.
19
+ """
20
+
21
+ def __init__(self, dataset, num_replicas, rank, ratio=1):
22
+ self.dataset = dataset
23
+ self.num_replicas = num_replicas
24
+ self.rank = rank
25
+ self.epoch = 0
26
+ self.num_samples = math.ceil(len(self.dataset) * ratio / self.num_replicas)
27
+ self.total_size = self.num_samples * self.num_replicas
28
+
29
+ def __iter__(self):
30
+ # deterministically shuffle based on epoch
31
+ g = torch.Generator()
32
+ g.manual_seed(self.epoch)
33
+ indices = torch.randperm(self.total_size, generator=g).tolist()
34
+
35
+ dataset_size = len(self.dataset)
36
+ indices = [v % dataset_size for v in indices]
37
+
38
+ # subsample
39
+ indices = indices[self.rank:self.total_size:self.num_replicas]
40
+ assert len(indices) == self.num_samples
41
+
42
+ return iter(indices)
43
+
44
+ def __len__(self):
45
+ return self.num_samples
46
+
47
+ def set_epoch(self, epoch):
48
+ self.epoch = epoch
basicsr/data/data_util.py ADDED
@@ -0,0 +1,315 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import cv2
2
+ import numpy as np
3
+ import torch
4
+ from os import path as osp
5
+ from torch.nn import functional as F
6
+
7
+ from basicsr.data.transforms import mod_crop
8
+ from basicsr.utils import img2tensor, scandir
9
+
10
+
11
+ def read_img_seq(path, require_mod_crop=False, scale=1, return_imgname=False):
12
+ """Read a sequence of images from a given folder path.
13
+
14
+ Args:
15
+ path (list[str] | str): List of image paths or image folder path.
16
+ require_mod_crop (bool): Require mod crop for each image.
17
+ Default: False.
18
+ scale (int): Scale factor for mod_crop. Default: 1.
19
+ return_imgname(bool): Whether return image names. Default False.
20
+
21
+ Returns:
22
+ Tensor: size (t, c, h, w), RGB, [0, 1].
23
+ list[str]: Returned image name list.
24
+ """
25
+ if isinstance(path, list):
26
+ img_paths = path
27
+ else:
28
+ img_paths = sorted(list(scandir(path, full_path=True)))
29
+ imgs = [cv2.imread(v).astype(np.float32) / 255. for v in img_paths]
30
+
31
+ if require_mod_crop:
32
+ imgs = [mod_crop(img, scale) for img in imgs]
33
+ imgs = img2tensor(imgs, bgr2rgb=True, float32=True)
34
+ imgs = torch.stack(imgs, dim=0)
35
+
36
+ if return_imgname:
37
+ imgnames = [osp.splitext(osp.basename(path))[0] for path in img_paths]
38
+ return imgs, imgnames
39
+ else:
40
+ return imgs
41
+
42
+
43
+ def generate_frame_indices(crt_idx, max_frame_num, num_frames, padding='reflection'):
44
+ """Generate an index list for reading `num_frames` frames from a sequence
45
+ of images.
46
+
47
+ Args:
48
+ crt_idx (int): Current center index.
49
+ max_frame_num (int): Max number of the sequence of images (from 1).
50
+ num_frames (int): Reading num_frames frames.
51
+ padding (str): Padding mode, one of
52
+ 'replicate' | 'reflection' | 'reflection_circle' | 'circle'
53
+ Examples: current_idx = 0, num_frames = 5
54
+ The generated frame indices under different padding mode:
55
+ replicate: [0, 0, 0, 1, 2]
56
+ reflection: [2, 1, 0, 1, 2]
57
+ reflection_circle: [4, 3, 0, 1, 2]
58
+ circle: [3, 4, 0, 1, 2]
59
+
60
+ Returns:
61
+ list[int]: A list of indices.
62
+ """
63
+ assert num_frames % 2 == 1, 'num_frames should be an odd number.'
64
+ assert padding in ('replicate', 'reflection', 'reflection_circle', 'circle'), f'Wrong padding mode: {padding}.'
65
+
66
+ max_frame_num = max_frame_num - 1 # start from 0
67
+ num_pad = num_frames // 2
68
+
69
+ indices = []
70
+ for i in range(crt_idx - num_pad, crt_idx + num_pad + 1):
71
+ if i < 0:
72
+ if padding == 'replicate':
73
+ pad_idx = 0
74
+ elif padding == 'reflection':
75
+ pad_idx = -i
76
+ elif padding == 'reflection_circle':
77
+ pad_idx = crt_idx + num_pad - i
78
+ else:
79
+ pad_idx = num_frames + i
80
+ elif i > max_frame_num:
81
+ if padding == 'replicate':
82
+ pad_idx = max_frame_num
83
+ elif padding == 'reflection':
84
+ pad_idx = max_frame_num * 2 - i
85
+ elif padding == 'reflection_circle':
86
+ pad_idx = (crt_idx - num_pad) - (i - max_frame_num)
87
+ else:
88
+ pad_idx = i - num_frames
89
+ else:
90
+ pad_idx = i
91
+ indices.append(pad_idx)
92
+ return indices
93
+
94
+
95
+ def paired_paths_from_lmdb(folders, keys):
96
+ """Generate paired paths from lmdb files.
97
+
98
+ Contents of lmdb. Taking the `lq.lmdb` for example, the file structure is:
99
+
100
+ ::
101
+
102
+ lq.lmdb
103
+ ├── data.mdb
104
+ ├── lock.mdb
105
+ ├── meta_info.txt
106
+
107
+ The data.mdb and lock.mdb are standard lmdb files and you can refer to
108
+ https://lmdb.readthedocs.io/en/release/ for more details.
109
+
110
+ The meta_info.txt is a specified txt file to record the meta information
111
+ of our datasets. It will be automatically created when preparing
112
+ datasets by our provided dataset tools.
113
+ Each line in the txt file records
114
+ 1)image name (with extension),
115
+ 2)image shape,
116
+ 3)compression level, separated by a white space.
117
+ Example: `baboon.png (120,125,3) 1`
118
+
119
+ We use the image name without extension as the lmdb key.
120
+ Note that we use the same key for the corresponding lq and gt images.
121
+
122
+ Args:
123
+ folders (list[str]): A list of folder path. The order of list should
124
+ be [input_folder, gt_folder].
125
+ keys (list[str]): A list of keys identifying folders. The order should
126
+ be in consistent with folders, e.g., ['lq', 'gt'].
127
+ Note that this key is different from lmdb keys.
128
+
129
+ Returns:
130
+ list[str]: Returned path list.
131
+ """
132
+ assert len(folders) == 2, ('The len of folders should be 2 with [input_folder, gt_folder]. '
133
+ f'But got {len(folders)}')
134
+ assert len(keys) == 2, f'The len of keys should be 2 with [input_key, gt_key]. But got {len(keys)}'
135
+ input_folder, gt_folder = folders
136
+ input_key, gt_key = keys
137
+
138
+ if not (input_folder.endswith('.lmdb') and gt_folder.endswith('.lmdb')):
139
+ raise ValueError(f'{input_key} folder and {gt_key} folder should both in lmdb '
140
+ f'formats. But received {input_key}: {input_folder}; '
141
+ f'{gt_key}: {gt_folder}')
142
+ # ensure that the two meta_info files are the same
143
+ with open(osp.join(input_folder, 'meta_info.txt')) as fin:
144
+ input_lmdb_keys = [line.split('.')[0] for line in fin]
145
+ with open(osp.join(gt_folder, 'meta_info.txt')) as fin:
146
+ gt_lmdb_keys = [line.split('.')[0] for line in fin]
147
+ if set(input_lmdb_keys) != set(gt_lmdb_keys):
148
+ raise ValueError(f'Keys in {input_key}_folder and {gt_key}_folder are different.')
149
+ else:
150
+ paths = []
151
+ for lmdb_key in sorted(input_lmdb_keys):
152
+ paths.append(dict([(f'{input_key}_path', lmdb_key), (f'{gt_key}_path', lmdb_key)]))
153
+ return paths
154
+
155
+
156
+ def paired_paths_from_meta_info_file(folders, keys, meta_info_file, filename_tmpl):
157
+ """Generate paired paths from an meta information file.
158
+
159
+ Each line in the meta information file contains the image names and
160
+ image shape (usually for gt), separated by a white space.
161
+
162
+ Example of an meta information file:
163
+ ```
164
+ 0001_s001.png (480,480,3)
165
+ 0001_s002.png (480,480,3)
166
+ ```
167
+
168
+ Args:
169
+ folders (list[str]): A list of folder path. The order of list should
170
+ be [input_folder, gt_folder].
171
+ keys (list[str]): A list of keys identifying folders. The order should
172
+ be in consistent with folders, e.g., ['lq', 'gt'].
173
+ meta_info_file (str): Path to the meta information file.
174
+ filename_tmpl (str): Template for each filename. Note that the
175
+ template excludes the file extension. Usually the filename_tmpl is
176
+ for files in the input folder.
177
+
178
+ Returns:
179
+ list[str]: Returned path list.
180
+ """
181
+ assert len(folders) == 2, ('The len of folders should be 2 with [input_folder, gt_folder]. '
182
+ f'But got {len(folders)}')
183
+ assert len(keys) == 2, f'The len of keys should be 2 with [input_key, gt_key]. But got {len(keys)}'
184
+ input_folder, gt_folder = folders
185
+ input_key, gt_key = keys
186
+
187
+ with open(meta_info_file, 'r') as fin:
188
+ gt_names = [line.strip().split(' ')[0] for line in fin]
189
+
190
+ paths = []
191
+ for gt_name in gt_names:
192
+ basename, ext = osp.splitext(osp.basename(gt_name))
193
+ input_name = f'{filename_tmpl.format(basename)}{ext}'
194
+ input_path = osp.join(input_folder, input_name)
195
+ gt_path = osp.join(gt_folder, gt_name)
196
+ paths.append(dict([(f'{input_key}_path', input_path), (f'{gt_key}_path', gt_path)]))
197
+ return paths
198
+
199
+
200
+ def paired_paths_from_folder(folders, keys, filename_tmpl):
201
+ """Generate paired paths from folders.
202
+
203
+ Args:
204
+ folders (list[str]): A list of folder path. The order of list should
205
+ be [input_folder, gt_folder].
206
+ keys (list[str]): A list of keys identifying folders. The order should
207
+ be in consistent with folders, e.g., ['lq', 'gt'].
208
+ filename_tmpl (str): Template for each filename. Note that the
209
+ template excludes the file extension. Usually the filename_tmpl is
210
+ for files in the input folder.
211
+
212
+ Returns:
213
+ list[str]: Returned path list.
214
+ """
215
+ assert len(folders) == 2, ('The len of folders should be 2 with [input_folder, gt_folder]. '
216
+ f'But got {len(folders)}')
217
+ assert len(keys) == 2, f'The len of keys should be 2 with [input_key, gt_key]. But got {len(keys)}'
218
+ input_folder, gt_folder = folders
219
+ input_key, gt_key = keys
220
+
221
+ input_paths = list(scandir(input_folder))
222
+ gt_paths = list(scandir(gt_folder))
223
+ assert len(input_paths) == len(gt_paths), (f'{input_key} and {gt_key} datasets have different number of images: '
224
+ f'{len(input_paths)}, {len(gt_paths)}.')
225
+ paths = []
226
+ for gt_path in gt_paths:
227
+ basename, ext = osp.splitext(osp.basename(gt_path))
228
+ input_name = f'{filename_tmpl.format(basename)}{ext}'
229
+ input_path = osp.join(input_folder, input_name)
230
+ assert input_name in input_paths, f'{input_name} is not in {input_key}_paths.'
231
+ gt_path = osp.join(gt_folder, gt_path)
232
+ paths.append(dict([(f'{input_key}_path', input_path), (f'{gt_key}_path', gt_path)]))
233
+ return paths
234
+
235
+
236
+ def paths_from_folder(folder):
237
+ """Generate paths from folder.
238
+
239
+ Args:
240
+ folder (str): Folder path.
241
+
242
+ Returns:
243
+ list[str]: Returned path list.
244
+ """
245
+
246
+ paths = list(scandir(folder))
247
+ paths = [osp.join(folder, path) for path in paths]
248
+ return paths
249
+
250
+
251
+ def paths_from_lmdb(folder):
252
+ """Generate paths from lmdb.
253
+
254
+ Args:
255
+ folder (str): Folder path.
256
+
257
+ Returns:
258
+ list[str]: Returned path list.
259
+ """
260
+ if not folder.endswith('.lmdb'):
261
+ raise ValueError(f'Folder {folder}folder should in lmdb format.')
262
+ with open(osp.join(folder, 'meta_info.txt')) as fin:
263
+ paths = [line.split('.')[0] for line in fin]
264
+ return paths
265
+
266
+
267
+ def generate_gaussian_kernel(kernel_size=13, sigma=1.6):
268
+ """Generate Gaussian kernel used in `duf_downsample`.
269
+
270
+ Args:
271
+ kernel_size (int): Kernel size. Default: 13.
272
+ sigma (float): Sigma of the Gaussian kernel. Default: 1.6.
273
+
274
+ Returns:
275
+ np.array: The Gaussian kernel.
276
+ """
277
+ from scipy.ndimage import filters as filters
278
+ kernel = np.zeros((kernel_size, kernel_size))
279
+ # set element at the middle to one, a dirac delta
280
+ kernel[kernel_size // 2, kernel_size // 2] = 1
281
+ # gaussian-smooth the dirac, resulting in a gaussian filter
282
+ return filters.gaussian_filter(kernel, sigma)
283
+
284
+
285
+ def duf_downsample(x, kernel_size=13, scale=4):
286
+ """Downsamping with Gaussian kernel used in the DUF official code.
287
+
288
+ Args:
289
+ x (Tensor): Frames to be downsampled, with shape (b, t, c, h, w).
290
+ kernel_size (int): Kernel size. Default: 13.
291
+ scale (int): Downsampling factor. Supported scale: (2, 3, 4).
292
+ Default: 4.
293
+
294
+ Returns:
295
+ Tensor: DUF downsampled frames.
296
+ """
297
+ assert scale in (2, 3, 4), f'Only support scale (2, 3, 4), but got {scale}.'
298
+
299
+ squeeze_flag = False
300
+ if x.ndim == 4:
301
+ squeeze_flag = True
302
+ x = x.unsqueeze(0)
303
+ b, t, c, h, w = x.size()
304
+ x = x.view(-1, 1, h, w)
305
+ pad_w, pad_h = kernel_size // 2 + scale * 2, kernel_size // 2 + scale * 2
306
+ x = F.pad(x, (pad_w, pad_w, pad_h, pad_h), 'reflect')
307
+
308
+ gaussian_filter = generate_gaussian_kernel(kernel_size, 0.4 * scale)
309
+ gaussian_filter = torch.from_numpy(gaussian_filter).type_as(x).unsqueeze(0).unsqueeze(0)
310
+ x = F.conv2d(x, gaussian_filter, stride=scale)
311
+ x = x[:, :, 2:-2, 2:-2]
312
+ x = x.view(b, t, c, x.size(2), x.size(3))
313
+ if squeeze_flag:
314
+ x = x.squeeze(0)
315
+ return x
basicsr/data/degradations.py ADDED
@@ -0,0 +1,765 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import cv2
2
+ import math
3
+ import numpy as np
4
+ import random
5
+ import torch
6
+ from scipy import special
7
+ from scipy.stats import multivariate_normal
8
+ # from torchvision.transforms.functional_tensor import rgb_to_grayscale
9
+ from torchvision.transforms.functional import rgb_to_grayscale
10
+
11
+ # -------------------------------------------------------------------- #
12
+ # --------------------------- blur kernels --------------------------- #
13
+ # -------------------------------------------------------------------- #
14
+
15
+
16
+ # --------------------------- util functions --------------------------- #
17
+ def sigma_matrix2(sig_x, sig_y, theta):
18
+ """Calculate the rotated sigma matrix (two dimensional matrix).
19
+
20
+ Args:
21
+ sig_x (float):
22
+ sig_y (float):
23
+ theta (float): Radian measurement.
24
+
25
+ Returns:
26
+ ndarray: Rotated sigma matrix.
27
+ """
28
+ d_matrix = np.array([[sig_x**2, 0], [0, sig_y**2]])
29
+ u_matrix = np.array([[np.cos(theta), -np.sin(theta)], [np.sin(theta), np.cos(theta)]])
30
+ return np.dot(u_matrix, np.dot(d_matrix, u_matrix.T))
31
+
32
+
33
+ def mesh_grid(kernel_size):
34
+ """Generate the mesh grid, centering at zero.
35
+
36
+ Args:
37
+ kernel_size (int):
38
+
39
+ Returns:
40
+ xy (ndarray): with the shape (kernel_size, kernel_size, 2)
41
+ xx (ndarray): with the shape (kernel_size, kernel_size)
42
+ yy (ndarray): with the shape (kernel_size, kernel_size)
43
+ """
44
+ ax = np.arange(-kernel_size // 2 + 1., kernel_size // 2 + 1.)
45
+ xx, yy = np.meshgrid(ax, ax)
46
+ xy = np.hstack((xx.reshape((kernel_size * kernel_size, 1)), yy.reshape(kernel_size * kernel_size,
47
+ 1))).reshape(kernel_size, kernel_size, 2)
48
+ return xy, xx, yy
49
+
50
+
51
+ def pdf2(sigma_matrix, grid):
52
+ """Calculate PDF of the bivariate Gaussian distribution.
53
+
54
+ Args:
55
+ sigma_matrix (ndarray): with the shape (2, 2)
56
+ grid (ndarray): generated by :func:`mesh_grid`,
57
+ with the shape (K, K, 2), K is the kernel size.
58
+
59
+ Returns:
60
+ kernel (ndarrray): un-normalized kernel.
61
+ """
62
+ inverse_sigma = np.linalg.inv(sigma_matrix)
63
+ kernel = np.exp(-0.5 * np.sum(np.dot(grid, inverse_sigma) * grid, 2))
64
+ return kernel
65
+
66
+
67
+ def cdf2(d_matrix, grid):
68
+ """Calculate the CDF of the standard bivariate Gaussian distribution.
69
+ Used in skewed Gaussian distribution.
70
+
71
+ Args:
72
+ d_matrix (ndarrasy): skew matrix.
73
+ grid (ndarray): generated by :func:`mesh_grid`,
74
+ with the shape (K, K, 2), K is the kernel size.
75
+
76
+ Returns:
77
+ cdf (ndarray): skewed cdf.
78
+ """
79
+ rv = multivariate_normal([0, 0], [[1, 0], [0, 1]])
80
+ grid = np.dot(grid, d_matrix)
81
+ cdf = rv.cdf(grid)
82
+ return cdf
83
+
84
+
85
+ def bivariate_Gaussian(kernel_size, sig_x, sig_y, theta, grid=None, isotropic=True):
86
+ """Generate a bivariate isotropic or anisotropic Gaussian kernel.
87
+
88
+ In the isotropic mode, only `sig_x` is used. `sig_y` and `theta` is ignored.
89
+
90
+ Args:
91
+ kernel_size (int):
92
+ sig_x (float):
93
+ sig_y (float):
94
+ theta (float): Radian measurement.
95
+ grid (ndarray, optional): generated by :func:`mesh_grid`,
96
+ with the shape (K, K, 2), K is the kernel size. Default: None
97
+ isotropic (bool):
98
+
99
+ Returns:
100
+ kernel (ndarray): normalized kernel.
101
+ """
102
+ if grid is None:
103
+ grid, _, _ = mesh_grid(kernel_size)
104
+ if isotropic:
105
+ sigma_matrix = np.array([[sig_x**2, 0], [0, sig_x**2]])
106
+ else:
107
+ sigma_matrix = sigma_matrix2(sig_x, sig_y, theta)
108
+ kernel = pdf2(sigma_matrix, grid)
109
+ kernel = kernel / np.sum(kernel)
110
+ return kernel
111
+
112
+
113
+ def bivariate_generalized_Gaussian(kernel_size, sig_x, sig_y, theta, beta, grid=None, isotropic=True):
114
+ """Generate a bivariate generalized Gaussian kernel.
115
+
116
+ ``Paper: Parameter Estimation For Multivariate Generalized Gaussian Distributions``
117
+
118
+ In the isotropic mode, only `sig_x` is used. `sig_y` and `theta` is ignored.
119
+
120
+ Args:
121
+ kernel_size (int):
122
+ sig_x (float):
123
+ sig_y (float):
124
+ theta (float): Radian measurement.
125
+ beta (float): shape parameter, beta = 1 is the normal distribution.
126
+ grid (ndarray, optional): generated by :func:`mesh_grid`,
127
+ with the shape (K, K, 2), K is the kernel size. Default: None
128
+
129
+ Returns:
130
+ kernel (ndarray): normalized kernel.
131
+ """
132
+ if grid is None:
133
+ grid, _, _ = mesh_grid(kernel_size)
134
+ if isotropic:
135
+ sigma_matrix = np.array([[sig_x**2, 0], [0, sig_x**2]])
136
+ else:
137
+ sigma_matrix = sigma_matrix2(sig_x, sig_y, theta)
138
+ inverse_sigma = np.linalg.inv(sigma_matrix)
139
+ kernel = np.exp(-0.5 * np.power(np.sum(np.dot(grid, inverse_sigma) * grid, 2), beta))
140
+ kernel = kernel / np.sum(kernel)
141
+ return kernel
142
+
143
+
144
+ def bivariate_plateau(kernel_size, sig_x, sig_y, theta, beta, grid=None, isotropic=True):
145
+ """Generate a plateau-like anisotropic kernel.
146
+
147
+ 1 / (1+x^(beta))
148
+
149
+ Reference: https://stats.stackexchange.com/questions/203629/is-there-a-plateau-shaped-distribution
150
+
151
+ In the isotropic mode, only `sig_x` is used. `sig_y` and `theta` is ignored.
152
+
153
+ Args:
154
+ kernel_size (int):
155
+ sig_x (float):
156
+ sig_y (float):
157
+ theta (float): Radian measurement.
158
+ beta (float): shape parameter, beta = 1 is the normal distribution.
159
+ grid (ndarray, optional): generated by :func:`mesh_grid`,
160
+ with the shape (K, K, 2), K is the kernel size. Default: None
161
+
162
+ Returns:
163
+ kernel (ndarray): normalized kernel.
164
+ """
165
+ if grid is None:
166
+ grid, _, _ = mesh_grid(kernel_size)
167
+ if isotropic:
168
+ sigma_matrix = np.array([[sig_x**2, 0], [0, sig_x**2]])
169
+ else:
170
+ sigma_matrix = sigma_matrix2(sig_x, sig_y, theta)
171
+ inverse_sigma = np.linalg.inv(sigma_matrix)
172
+ kernel = np.reciprocal(np.power(np.sum(np.dot(grid, inverse_sigma) * grid, 2), beta) + 1)
173
+ kernel = kernel / np.sum(kernel)
174
+ return kernel
175
+
176
+
177
+ def random_bivariate_Gaussian(kernel_size,
178
+ sigma_x_range,
179
+ sigma_y_range,
180
+ rotation_range,
181
+ noise_range=None,
182
+ isotropic=True):
183
+ """Randomly generate bivariate isotropic or anisotropic Gaussian kernels.
184
+
185
+ In the isotropic mode, only `sigma_x_range` is used. `sigma_y_range` and `rotation_range` is ignored.
186
+
187
+ Args:
188
+ kernel_size (int):
189
+ sigma_x_range (tuple): [0.6, 5]
190
+ sigma_y_range (tuple): [0.6, 5]
191
+ rotation range (tuple): [-math.pi, math.pi]
192
+ noise_range(tuple, optional): multiplicative kernel noise,
193
+ [0.75, 1.25]. Default: None
194
+
195
+ Returns:
196
+ kernel (ndarray):
197
+ """
198
+ assert kernel_size % 2 == 1, 'Kernel size must be an odd number.'
199
+ assert sigma_x_range[0] < sigma_x_range[1], 'Wrong sigma_x_range.'
200
+ sigma_x = np.random.uniform(sigma_x_range[0], sigma_x_range[1])
201
+ if isotropic is False:
202
+ assert sigma_y_range[0] < sigma_y_range[1], 'Wrong sigma_y_range.'
203
+ assert rotation_range[0] < rotation_range[1], 'Wrong rotation_range.'
204
+ sigma_y = np.random.uniform(sigma_y_range[0], sigma_y_range[1])
205
+ rotation = np.random.uniform(rotation_range[0], rotation_range[1])
206
+ else:
207
+ sigma_y = sigma_x
208
+ rotation = 0
209
+
210
+ kernel = bivariate_Gaussian(kernel_size, sigma_x, sigma_y, rotation, isotropic=isotropic)
211
+
212
+ # add multiplicative noise
213
+ if noise_range is not None:
214
+ assert noise_range[0] < noise_range[1], 'Wrong noise range.'
215
+ noise = np.random.uniform(noise_range[0], noise_range[1], size=kernel.shape)
216
+ kernel = kernel * noise
217
+ kernel = kernel / np.sum(kernel)
218
+ return kernel
219
+
220
+
221
+ def random_bivariate_generalized_Gaussian(kernel_size,
222
+ sigma_x_range,
223
+ sigma_y_range,
224
+ rotation_range,
225
+ beta_range,
226
+ noise_range=None,
227
+ isotropic=True):
228
+ """Randomly generate bivariate generalized Gaussian kernels.
229
+
230
+ In the isotropic mode, only `sigma_x_range` is used. `sigma_y_range` and `rotation_range` is ignored.
231
+
232
+ Args:
233
+ kernel_size (int):
234
+ sigma_x_range (tuple): [0.6, 5]
235
+ sigma_y_range (tuple): [0.6, 5]
236
+ rotation range (tuple): [-math.pi, math.pi]
237
+ beta_range (tuple): [0.5, 8]
238
+ noise_range(tuple, optional): multiplicative kernel noise,
239
+ [0.75, 1.25]. Default: None
240
+
241
+ Returns:
242
+ kernel (ndarray):
243
+ """
244
+ assert kernel_size % 2 == 1, 'Kernel size must be an odd number.'
245
+ assert sigma_x_range[0] < sigma_x_range[1], 'Wrong sigma_x_range.'
246
+ sigma_x = np.random.uniform(sigma_x_range[0], sigma_x_range[1])
247
+ if isotropic is False:
248
+ assert sigma_y_range[0] < sigma_y_range[1], 'Wrong sigma_y_range.'
249
+ assert rotation_range[0] < rotation_range[1], 'Wrong rotation_range.'
250
+ sigma_y = np.random.uniform(sigma_y_range[0], sigma_y_range[1])
251
+ rotation = np.random.uniform(rotation_range[0], rotation_range[1])
252
+ else:
253
+ sigma_y = sigma_x
254
+ rotation = 0
255
+
256
+ # assume beta_range[0] < 1 < beta_range[1]
257
+ if np.random.uniform() < 0.5:
258
+ beta = np.random.uniform(beta_range[0], 1)
259
+ else:
260
+ beta = np.random.uniform(1, beta_range[1])
261
+
262
+ kernel = bivariate_generalized_Gaussian(kernel_size, sigma_x, sigma_y, rotation, beta, isotropic=isotropic)
263
+
264
+ # add multiplicative noise
265
+ if noise_range is not None:
266
+ assert noise_range[0] < noise_range[1], 'Wrong noise range.'
267
+ noise = np.random.uniform(noise_range[0], noise_range[1], size=kernel.shape)
268
+ kernel = kernel * noise
269
+ kernel = kernel / np.sum(kernel)
270
+ return kernel
271
+
272
+
273
+ def random_bivariate_plateau(kernel_size,
274
+ sigma_x_range,
275
+ sigma_y_range,
276
+ rotation_range,
277
+ beta_range,
278
+ noise_range=None,
279
+ isotropic=True):
280
+ """Randomly generate bivariate plateau kernels.
281
+
282
+ In the isotropic mode, only `sigma_x_range` is used. `sigma_y_range` and `rotation_range` is ignored.
283
+
284
+ Args:
285
+ kernel_size (int):
286
+ sigma_x_range (tuple): [0.6, 5]
287
+ sigma_y_range (tuple): [0.6, 5]
288
+ rotation range (tuple): [-math.pi/2, math.pi/2]
289
+ beta_range (tuple): [1, 4]
290
+ noise_range(tuple, optional): multiplicative kernel noise,
291
+ [0.75, 1.25]. Default: None
292
+
293
+ Returns:
294
+ kernel (ndarray):
295
+ """
296
+ assert kernel_size % 2 == 1, 'Kernel size must be an odd number.'
297
+ assert sigma_x_range[0] < sigma_x_range[1], 'Wrong sigma_x_range.'
298
+ sigma_x = np.random.uniform(sigma_x_range[0], sigma_x_range[1])
299
+ if isotropic is False:
300
+ assert sigma_y_range[0] < sigma_y_range[1], 'Wrong sigma_y_range.'
301
+ assert rotation_range[0] < rotation_range[1], 'Wrong rotation_range.'
302
+ sigma_y = np.random.uniform(sigma_y_range[0], sigma_y_range[1])
303
+ rotation = np.random.uniform(rotation_range[0], rotation_range[1])
304
+ else:
305
+ sigma_y = sigma_x
306
+ rotation = 0
307
+
308
+ # TODO: this may be not proper
309
+ if np.random.uniform() < 0.5:
310
+ beta = np.random.uniform(beta_range[0], 1)
311
+ else:
312
+ beta = np.random.uniform(1, beta_range[1])
313
+
314
+ kernel = bivariate_plateau(kernel_size, sigma_x, sigma_y, rotation, beta, isotropic=isotropic)
315
+ # add multiplicative noise
316
+ if noise_range is not None:
317
+ assert noise_range[0] < noise_range[1], 'Wrong noise range.'
318
+ noise = np.random.uniform(noise_range[0], noise_range[1], size=kernel.shape)
319
+ kernel = kernel * noise
320
+ kernel = kernel / np.sum(kernel)
321
+
322
+ return kernel
323
+
324
+
325
+ def random_mixed_kernels(kernel_list,
326
+ kernel_prob,
327
+ kernel_size=21,
328
+ sigma_x_range=(0.6, 5),
329
+ sigma_y_range=(0.6, 5),
330
+ rotation_range=(-math.pi, math.pi),
331
+ betag_range=(0.5, 8),
332
+ betap_range=(0.5, 8),
333
+ noise_range=None):
334
+ """Randomly generate mixed kernels.
335
+
336
+ Args:
337
+ kernel_list (tuple): a list name of kernel types,
338
+ support ['iso', 'aniso', 'skew', 'generalized', 'plateau_iso',
339
+ 'plateau_aniso']
340
+ kernel_prob (tuple): corresponding kernel probability for each
341
+ kernel type
342
+ kernel_size (int):
343
+ sigma_x_range (tuple): [0.6, 5]
344
+ sigma_y_range (tuple): [0.6, 5]
345
+ rotation range (tuple): [-math.pi, math.pi]
346
+ beta_range (tuple): [0.5, 8]
347
+ noise_range(tuple, optional): multiplicative kernel noise,
348
+ [0.75, 1.25]. Default: None
349
+
350
+ Returns:
351
+ kernel (ndarray):
352
+ """
353
+ kernel_type = random.choices(kernel_list, kernel_prob)[0]
354
+ if kernel_type == 'iso':
355
+ kernel = random_bivariate_Gaussian(
356
+ kernel_size, sigma_x_range, sigma_y_range, rotation_range, noise_range=noise_range, isotropic=True)
357
+ elif kernel_type == 'aniso':
358
+ kernel = random_bivariate_Gaussian(
359
+ kernel_size, sigma_x_range, sigma_y_range, rotation_range, noise_range=noise_range, isotropic=False)
360
+ elif kernel_type == 'generalized_iso':
361
+ kernel = random_bivariate_generalized_Gaussian(
362
+ kernel_size,
363
+ sigma_x_range,
364
+ sigma_y_range,
365
+ rotation_range,
366
+ betag_range,
367
+ noise_range=noise_range,
368
+ isotropic=True)
369
+ elif kernel_type == 'generalized_aniso':
370
+ kernel = random_bivariate_generalized_Gaussian(
371
+ kernel_size,
372
+ sigma_x_range,
373
+ sigma_y_range,
374
+ rotation_range,
375
+ betag_range,
376
+ noise_range=noise_range,
377
+ isotropic=False)
378
+ elif kernel_type == 'plateau_iso':
379
+ kernel = random_bivariate_plateau(
380
+ kernel_size, sigma_x_range, sigma_y_range, rotation_range, betap_range, noise_range=None, isotropic=True)
381
+ elif kernel_type == 'plateau_aniso':
382
+ kernel = random_bivariate_plateau(
383
+ kernel_size, sigma_x_range, sigma_y_range, rotation_range, betap_range, noise_range=None, isotropic=False)
384
+ return kernel
385
+
386
+
387
+ np.seterr(divide='ignore', invalid='ignore')
388
+
389
+
390
+ def circular_lowpass_kernel(cutoff, kernel_size, pad_to=0):
391
+ """2D sinc filter
392
+
393
+ Reference: https://dsp.stackexchange.com/questions/58301/2-d-circularly-symmetric-low-pass-filter
394
+
395
+ Args:
396
+ cutoff (float): cutoff frequency in radians (pi is max)
397
+ kernel_size (int): horizontal and vertical size, must be odd.
398
+ pad_to (int): pad kernel size to desired size, must be odd or zero.
399
+ """
400
+ assert kernel_size % 2 == 1, 'Kernel size must be an odd number.'
401
+ kernel = np.fromfunction(
402
+ lambda x, y: cutoff * special.j1(cutoff * np.sqrt(
403
+ (x - (kernel_size - 1) / 2)**2 + (y - (kernel_size - 1) / 2)**2)) / (2 * np.pi * np.sqrt(
404
+ (x - (kernel_size - 1) / 2)**2 + (y - (kernel_size - 1) / 2)**2)), [kernel_size, kernel_size])
405
+ kernel[(kernel_size - 1) // 2, (kernel_size - 1) // 2] = cutoff**2 / (4 * np.pi)
406
+ kernel = kernel / np.sum(kernel)
407
+ if pad_to > kernel_size:
408
+ pad_size = (pad_to - kernel_size) // 2
409
+ kernel = np.pad(kernel, ((pad_size, pad_size), (pad_size, pad_size)))
410
+ return kernel
411
+
412
+
413
+ # ------------------------------------------------------------- #
414
+ # --------------------------- noise --------------------------- #
415
+ # ------------------------------------------------------------- #
416
+
417
+ # ----------------------- Gaussian Noise ----------------------- #
418
+
419
+
420
+ def generate_gaussian_noise(img, sigma=10, gray_noise=False):
421
+ """Generate Gaussian noise.
422
+
423
+ Args:
424
+ img (Numpy array): Input image, shape (h, w, c), range [0, 1], float32.
425
+ sigma (float): Noise scale (measured in range 255). Default: 10.
426
+
427
+ Returns:
428
+ (Numpy array): Returned noisy image, shape (h, w, c), range[0, 1],
429
+ float32.
430
+ """
431
+ if gray_noise:
432
+ noise = np.float32(np.random.randn(*(img.shape[0:2]))) * sigma / 255.
433
+ noise = np.expand_dims(noise, axis=2).repeat(3, axis=2)
434
+ else:
435
+ noise = np.float32(np.random.randn(*(img.shape))) * sigma / 255.
436
+ return noise
437
+
438
+
439
+ def add_gaussian_noise(img, sigma=10, clip=True, rounds=False, gray_noise=False):
440
+ """Add Gaussian noise.
441
+
442
+ Args:
443
+ img (Numpy array): Input image, shape (h, w, c), range [0, 1], float32.
444
+ sigma (float): Noise scale (measured in range 255). Default: 10.
445
+
446
+ Returns:
447
+ (Numpy array): Returned noisy image, shape (h, w, c), range[0, 1],
448
+ float32.
449
+ """
450
+ noise = generate_gaussian_noise(img, sigma, gray_noise)
451
+ out = img + noise
452
+ if clip and rounds:
453
+ out = np.clip((out * 255.0).round(), 0, 255) / 255.
454
+ elif clip:
455
+ out = np.clip(out, 0, 1)
456
+ elif rounds:
457
+ out = (out * 255.0).round() / 255.
458
+ return out
459
+
460
+
461
+ def generate_gaussian_noise_pt(img, sigma=10, gray_noise=0):
462
+ """Add Gaussian noise (PyTorch version).
463
+
464
+ Args:
465
+ img (Tensor): Shape (b, c, h, w), range[0, 1], float32.
466
+ scale (float | Tensor): Noise scale. Default: 1.0.
467
+
468
+ Returns:
469
+ (Tensor): Returned noisy image, shape (b, c, h, w), range[0, 1],
470
+ float32.
471
+ """
472
+ b, _, h, w = img.size()
473
+ if not isinstance(sigma, (float, int)):
474
+ sigma = sigma.view(img.size(0), 1, 1, 1)
475
+ if isinstance(gray_noise, (float, int)):
476
+ cal_gray_noise = gray_noise > 0
477
+ else:
478
+ gray_noise = gray_noise.view(b, 1, 1, 1)
479
+ cal_gray_noise = torch.sum(gray_noise) > 0
480
+
481
+ if cal_gray_noise:
482
+ noise_gray = torch.randn(*img.size()[2:4], dtype=img.dtype, device=img.device) * sigma / 255.
483
+ noise_gray = noise_gray.view(b, 1, h, w)
484
+
485
+ # always calculate color noise
486
+ noise = torch.randn(*img.size(), dtype=img.dtype, device=img.device) * sigma / 255.
487
+
488
+ if cal_gray_noise:
489
+ noise = noise * (1 - gray_noise) + noise_gray * gray_noise
490
+ return noise
491
+
492
+
493
+ def add_gaussian_noise_pt(img, sigma=10, gray_noise=0, clip=True, rounds=False):
494
+ """Add Gaussian noise (PyTorch version).
495
+
496
+ Args:
497
+ img (Tensor): Shape (b, c, h, w), range[0, 1], float32.
498
+ scale (float | Tensor): Noise scale. Default: 1.0.
499
+
500
+ Returns:
501
+ (Tensor): Returned noisy image, shape (b, c, h, w), range[0, 1],
502
+ float32.
503
+ """
504
+ noise = generate_gaussian_noise_pt(img, sigma, gray_noise)
505
+ out = img + noise
506
+ if clip and rounds:
507
+ out = torch.clamp((out * 255.0).round(), 0, 255) / 255.
508
+ elif clip:
509
+ out = torch.clamp(out, 0, 1)
510
+ elif rounds:
511
+ out = (out * 255.0).round() / 255.
512
+ return out
513
+
514
+
515
+ # ----------------------- Random Gaussian Noise ----------------------- #
516
+ def random_generate_gaussian_noise(img, sigma_range=(0, 10), gray_prob=0):
517
+ sigma = np.random.uniform(sigma_range[0], sigma_range[1])
518
+ if np.random.uniform() < gray_prob:
519
+ gray_noise = True
520
+ else:
521
+ gray_noise = False
522
+ return generate_gaussian_noise(img, sigma, gray_noise)
523
+
524
+
525
+ def random_add_gaussian_noise(img, sigma_range=(0, 1.0), gray_prob=0, clip=True, rounds=False):
526
+ noise = random_generate_gaussian_noise(img, sigma_range, gray_prob)
527
+ out = img + noise
528
+ if clip and rounds:
529
+ out = np.clip((out * 255.0).round(), 0, 255) / 255.
530
+ elif clip:
531
+ out = np.clip(out, 0, 1)
532
+ elif rounds:
533
+ out = (out * 255.0).round() / 255.
534
+ return out
535
+
536
+
537
+ def random_generate_gaussian_noise_pt(img, sigma_range=(0, 10), gray_prob=0):
538
+ sigma = torch.rand(
539
+ img.size(0), dtype=img.dtype, device=img.device) * (sigma_range[1] - sigma_range[0]) + sigma_range[0]
540
+ gray_noise = torch.rand(img.size(0), dtype=img.dtype, device=img.device)
541
+ gray_noise = (gray_noise < gray_prob).float()
542
+ return generate_gaussian_noise_pt(img, sigma, gray_noise)
543
+
544
+
545
+ def random_add_gaussian_noise_pt(img, sigma_range=(0, 1.0), gray_prob=0, clip=True, rounds=False):
546
+ noise = random_generate_gaussian_noise_pt(img, sigma_range, gray_prob)
547
+ out = img + noise
548
+ if clip and rounds:
549
+ out = torch.clamp((out * 255.0).round(), 0, 255) / 255.
550
+ elif clip:
551
+ out = torch.clamp(out, 0, 1)
552
+ elif rounds:
553
+ out = (out * 255.0).round() / 255.
554
+ return out
555
+
556
+
557
+ # ----------------------- Poisson (Shot) Noise ----------------------- #
558
+
559
+
560
+ def generate_poisson_noise(img, scale=1.0, gray_noise=False):
561
+ """Generate poisson noise.
562
+
563
+ Reference: https://github.com/scikit-image/scikit-image/blob/main/skimage/util/noise.py#L37-L219
564
+
565
+ Args:
566
+ img (Numpy array): Input image, shape (h, w, c), range [0, 1], float32.
567
+ scale (float): Noise scale. Default: 1.0.
568
+ gray_noise (bool): Whether generate gray noise. Default: False.
569
+
570
+ Returns:
571
+ (Numpy array): Returned noisy image, shape (h, w, c), range[0, 1],
572
+ float32.
573
+ """
574
+ if gray_noise:
575
+ img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
576
+ # round and clip image for counting vals correctly
577
+ img = np.clip((img * 255.0).round(), 0, 255) / 255.
578
+ vals = len(np.unique(img))
579
+ vals = 2**np.ceil(np.log2(vals))
580
+ out = np.float32(np.random.poisson(img * vals) / float(vals))
581
+ noise = out - img
582
+ if gray_noise:
583
+ noise = np.repeat(noise[:, :, np.newaxis], 3, axis=2)
584
+ return noise * scale
585
+
586
+
587
+ def add_poisson_noise(img, scale=1.0, clip=True, rounds=False, gray_noise=False):
588
+ """Add poisson noise.
589
+
590
+ Args:
591
+ img (Numpy array): Input image, shape (h, w, c), range [0, 1], float32.
592
+ scale (float): Noise scale. Default: 1.0.
593
+ gray_noise (bool): Whether generate gray noise. Default: False.
594
+
595
+ Returns:
596
+ (Numpy array): Returned noisy image, shape (h, w, c), range[0, 1],
597
+ float32.
598
+ """
599
+ noise = generate_poisson_noise(img, scale, gray_noise)
600
+ out = img + noise
601
+ if clip and rounds:
602
+ out = np.clip((out * 255.0).round(), 0, 255) / 255.
603
+ elif clip:
604
+ out = np.clip(out, 0, 1)
605
+ elif rounds:
606
+ out = (out * 255.0).round() / 255.
607
+ return out
608
+
609
+
610
+ def generate_poisson_noise_pt(img, scale=1.0, gray_noise=0):
611
+ """Generate a batch of poisson noise (PyTorch version)
612
+
613
+ Args:
614
+ img (Tensor): Input image, shape (b, c, h, w), range [0, 1], float32.
615
+ scale (float | Tensor): Noise scale. Number or Tensor with shape (b).
616
+ Default: 1.0.
617
+ gray_noise (float | Tensor): 0-1 number or Tensor with shape (b).
618
+ 0 for False, 1 for True. Default: 0.
619
+
620
+ Returns:
621
+ (Tensor): Returned noisy image, shape (b, c, h, w), range[0, 1],
622
+ float32.
623
+ """
624
+ b, _, h, w = img.size()
625
+ if isinstance(gray_noise, (float, int)):
626
+ cal_gray_noise = gray_noise > 0
627
+ else:
628
+ gray_noise = gray_noise.view(b, 1, 1, 1)
629
+ cal_gray_noise = torch.sum(gray_noise) > 0
630
+ if cal_gray_noise:
631
+ img_gray = rgb_to_grayscale(img, num_output_channels=1)
632
+ # round and clip image for counting vals correctly
633
+ img_gray = torch.clamp((img_gray * 255.0).round(), 0, 255) / 255.
634
+ # use for-loop to get the unique values for each sample
635
+ vals_list = [len(torch.unique(img_gray[i, :, :, :])) for i in range(b)]
636
+ vals_list = [2**np.ceil(np.log2(vals)) for vals in vals_list]
637
+ vals = img_gray.new_tensor(vals_list).view(b, 1, 1, 1)
638
+ out = torch.poisson(img_gray * vals) / vals
639
+ noise_gray = out - img_gray
640
+ noise_gray = noise_gray.expand(b, 3, h, w)
641
+
642
+ # always calculate color noise
643
+ # round and clip image for counting vals correctly
644
+ img = torch.clamp((img * 255.0).round(), 0, 255) / 255.
645
+ # use for-loop to get the unique values for each sample
646
+ vals_list = [len(torch.unique(img[i, :, :, :])) for i in range(b)]
647
+ vals_list = [2**np.ceil(np.log2(vals)) for vals in vals_list]
648
+ vals = img.new_tensor(vals_list).view(b, 1, 1, 1)
649
+ out = torch.poisson(img * vals) / vals
650
+ noise = out - img
651
+ if cal_gray_noise:
652
+ noise = noise * (1 - gray_noise) + noise_gray * gray_noise
653
+ if not isinstance(scale, (float, int)):
654
+ scale = scale.view(b, 1, 1, 1)
655
+ return noise * scale
656
+
657
+
658
+ def add_poisson_noise_pt(img, scale=1.0, clip=True, rounds=False, gray_noise=0):
659
+ """Add poisson noise to a batch of images (PyTorch version).
660
+
661
+ Args:
662
+ img (Tensor): Input image, shape (b, c, h, w), range [0, 1], float32.
663
+ scale (float | Tensor): Noise scale. Number or Tensor with shape (b).
664
+ Default: 1.0.
665
+ gray_noise (float | Tensor): 0-1 number or Tensor with shape (b).
666
+ 0 for False, 1 for True. Default: 0.
667
+
668
+ Returns:
669
+ (Tensor): Returned noisy image, shape (b, c, h, w), range[0, 1],
670
+ float32.
671
+ """
672
+ noise = generate_poisson_noise_pt(img, scale, gray_noise)
673
+ out = img + noise
674
+ if clip and rounds:
675
+ out = torch.clamp((out * 255.0).round(), 0, 255) / 255.
676
+ elif clip:
677
+ out = torch.clamp(out, 0, 1)
678
+ elif rounds:
679
+ out = (out * 255.0).round() / 255.
680
+ return out
681
+
682
+
683
+ # ----------------------- Random Poisson (Shot) Noise ----------------------- #
684
+
685
+
686
+ def random_generate_poisson_noise(img, scale_range=(0, 1.0), gray_prob=0):
687
+ scale = np.random.uniform(scale_range[0], scale_range[1])
688
+ if np.random.uniform() < gray_prob:
689
+ gray_noise = True
690
+ else:
691
+ gray_noise = False
692
+ return generate_poisson_noise(img, scale, gray_noise)
693
+
694
+
695
+ def random_add_poisson_noise(img, scale_range=(0, 1.0), gray_prob=0, clip=True, rounds=False):
696
+ noise = random_generate_poisson_noise(img, scale_range, gray_prob)
697
+ out = img + noise
698
+ if clip and rounds:
699
+ out = np.clip((out * 255.0).round(), 0, 255) / 255.
700
+ elif clip:
701
+ out = np.clip(out, 0, 1)
702
+ elif rounds:
703
+ out = (out * 255.0).round() / 255.
704
+ return out
705
+
706
+
707
+ def random_generate_poisson_noise_pt(img, scale_range=(0, 1.0), gray_prob=0):
708
+ scale = torch.rand(
709
+ img.size(0), dtype=img.dtype, device=img.device) * (scale_range[1] - scale_range[0]) + scale_range[0]
710
+ gray_noise = torch.rand(img.size(0), dtype=img.dtype, device=img.device)
711
+ gray_noise = (gray_noise < gray_prob).float()
712
+ return generate_poisson_noise_pt(img, scale, gray_noise)
713
+
714
+
715
+ def random_add_poisson_noise_pt(img, scale_range=(0, 1.0), gray_prob=0, clip=True, rounds=False):
716
+ noise = random_generate_poisson_noise_pt(img, scale_range, gray_prob)
717
+ out = img + noise
718
+ if clip and rounds:
719
+ out = torch.clamp((out * 255.0).round(), 0, 255) / 255.
720
+ elif clip:
721
+ out = torch.clamp(out, 0, 1)
722
+ elif rounds:
723
+ out = (out * 255.0).round() / 255.
724
+ return out
725
+
726
+
727
+ # ------------------------------------------------------------------------ #
728
+ # --------------------------- JPEG compression --------------------------- #
729
+ # ------------------------------------------------------------------------ #
730
+
731
+
732
+ def add_jpg_compression(img, quality=90):
733
+ """Add JPG compression artifacts.
734
+
735
+ Args:
736
+ img (Numpy array): Input image, shape (h, w, c), range [0, 1], float32.
737
+ quality (float): JPG compression quality. 0 for lowest quality, 100 for
738
+ best quality. Default: 90.
739
+
740
+ Returns:
741
+ (Numpy array): Returned image after JPG, shape (h, w, c), range[0, 1],
742
+ float32.
743
+ """
744
+ img = np.clip(img, 0, 1)
745
+ encode_param = [int(cv2.IMWRITE_JPEG_QUALITY), int(quality)]
746
+ _, encimg = cv2.imencode('.jpg', img * 255., encode_param)
747
+ img = np.float32(cv2.imdecode(encimg, 1)) / 255.
748
+ return img
749
+
750
+
751
+ def random_add_jpg_compression(img, quality_range=(90, 100)):
752
+ """Randomly add JPG compression artifacts.
753
+
754
+ Args:
755
+ img (Numpy array): Input image, shape (h, w, c), range [0, 1], float32.
756
+ quality_range (tuple[float] | list[float]): JPG compression quality
757
+ range. 0 for lowest quality, 100 for best quality.
758
+ Default: (90, 100).
759
+
760
+ Returns:
761
+ (Numpy array): Returned image after JPG, shape (h, w, c), range[0, 1],
762
+ float32.
763
+ """
764
+ quality = np.random.uniform(quality_range[0], quality_range[1])
765
+ return add_jpg_compression(img, quality)
basicsr/data/ffhq_dataset.py ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import random
2
+ import time
3
+ from os import path as osp
4
+ from torch.utils import data as data
5
+ from torchvision.transforms.functional import normalize
6
+
7
+ from basicsr.data.transforms import augment
8
+ from basicsr.utils import FileClient, get_root_logger, imfrombytes, img2tensor
9
+ from basicsr.utils.registry import DATASET_REGISTRY
10
+
11
+
12
+ @DATASET_REGISTRY.register()
13
+ class FFHQDataset(data.Dataset):
14
+ """FFHQ dataset for StyleGAN.
15
+
16
+ Args:
17
+ opt (dict): Config for train datasets. It contains the following keys:
18
+ dataroot_gt (str): Data root path for gt.
19
+ io_backend (dict): IO backend type and other kwarg.
20
+ mean (list | tuple): Image mean.
21
+ std (list | tuple): Image std.
22
+ use_hflip (bool): Whether to horizontally flip.
23
+
24
+ """
25
+
26
+ def __init__(self, opt):
27
+ super(FFHQDataset, self).__init__()
28
+ self.opt = opt
29
+ # file client (io backend)
30
+ self.file_client = None
31
+ self.io_backend_opt = opt['io_backend']
32
+
33
+ self.gt_folder = opt['dataroot_gt']
34
+ self.mean = opt['mean']
35
+ self.std = opt['std']
36
+
37
+ if self.io_backend_opt['type'] == 'lmdb':
38
+ self.io_backend_opt['db_paths'] = self.gt_folder
39
+ if not self.gt_folder.endswith('.lmdb'):
40
+ raise ValueError("'dataroot_gt' should end with '.lmdb', but received {self.gt_folder}")
41
+ with open(osp.join(self.gt_folder, 'meta_info.txt')) as fin:
42
+ self.paths = [line.split('.')[0] for line in fin]
43
+ else:
44
+ # FFHQ has 70000 images in total
45
+ self.paths = [osp.join(self.gt_folder, f'{v:08d}.png') for v in range(70000)]
46
+
47
+ def __getitem__(self, index):
48
+ if self.file_client is None:
49
+ self.file_client = FileClient(self.io_backend_opt.pop('type'), **self.io_backend_opt)
50
+
51
+ # load gt image
52
+ gt_path = self.paths[index]
53
+ # avoid errors caused by high latency in reading files
54
+ retry = 3
55
+ while retry > 0:
56
+ try:
57
+ img_bytes = self.file_client.get(gt_path)
58
+ except Exception as e:
59
+ logger = get_root_logger()
60
+ logger.warning(f'File client error: {e}, remaining retry times: {retry - 1}')
61
+ # change another file to read
62
+ index = random.randint(0, self.__len__())
63
+ gt_path = self.paths[index]
64
+ time.sleep(1) # sleep 1s for occasional server congestion
65
+ else:
66
+ break
67
+ finally:
68
+ retry -= 1
69
+ img_gt = imfrombytes(img_bytes, float32=True)
70
+
71
+ # random horizontal flip
72
+ img_gt = augment(img_gt, hflip=self.opt['use_hflip'], rotation=False)
73
+ # BGR to RGB, HWC to CHW, numpy to tensor
74
+ img_gt = img2tensor(img_gt, bgr2rgb=True, float32=True)
75
+ # normalize
76
+ normalize(img_gt, self.mean, self.std, inplace=True)
77
+ return {'gt': img_gt, 'gt_path': gt_path}
78
+
79
+ def __len__(self):
80
+ return len(self.paths)
basicsr/data/meta_info/meta_info_DIV2K800sub_GT.txt ADDED
The diff for this file is too large to render. See raw diff
 
basicsr/data/meta_info/meta_info_REDS4_test_GT.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ 000 100 (720,1280,3)
2
+ 011 100 (720,1280,3)
3
+ 015 100 (720,1280,3)
4
+ 020 100 (720,1280,3)
basicsr/data/meta_info/meta_info_REDS_GT.txt ADDED
@@ -0,0 +1,270 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 000 100 (720,1280,3)
2
+ 001 100 (720,1280,3)
3
+ 002 100 (720,1280,3)
4
+ 003 100 (720,1280,3)
5
+ 004 100 (720,1280,3)
6
+ 005 100 (720,1280,3)
7
+ 006 100 (720,1280,3)
8
+ 007 100 (720,1280,3)
9
+ 008 100 (720,1280,3)
10
+ 009 100 (720,1280,3)
11
+ 010 100 (720,1280,3)
12
+ 011 100 (720,1280,3)
13
+ 012 100 (720,1280,3)
14
+ 013 100 (720,1280,3)
15
+ 014 100 (720,1280,3)
16
+ 015 100 (720,1280,3)
17
+ 016 100 (720,1280,3)
18
+ 017 100 (720,1280,3)
19
+ 018 100 (720,1280,3)
20
+ 019 100 (720,1280,3)
21
+ 020 100 (720,1280,3)
22
+ 021 100 (720,1280,3)
23
+ 022 100 (720,1280,3)
24
+ 023 100 (720,1280,3)
25
+ 024 100 (720,1280,3)
26
+ 025 100 (720,1280,3)
27
+ 026 100 (720,1280,3)
28
+ 027 100 (720,1280,3)
29
+ 028 100 (720,1280,3)
30
+ 029 100 (720,1280,3)
31
+ 030 100 (720,1280,3)
32
+ 031 100 (720,1280,3)
33
+ 032 100 (720,1280,3)
34
+ 033 100 (720,1280,3)
35
+ 034 100 (720,1280,3)
36
+ 035 100 (720,1280,3)
37
+ 036 100 (720,1280,3)
38
+ 037 100 (720,1280,3)
39
+ 038 100 (720,1280,3)
40
+ 039 100 (720,1280,3)
41
+ 040 100 (720,1280,3)
42
+ 041 100 (720,1280,3)
43
+ 042 100 (720,1280,3)
44
+ 043 100 (720,1280,3)
45
+ 044 100 (720,1280,3)
46
+ 045 100 (720,1280,3)
47
+ 046 100 (720,1280,3)
48
+ 047 100 (720,1280,3)
49
+ 048 100 (720,1280,3)
50
+ 049 100 (720,1280,3)
51
+ 050 100 (720,1280,3)
52
+ 051 100 (720,1280,3)
53
+ 052 100 (720,1280,3)
54
+ 053 100 (720,1280,3)
55
+ 054 100 (720,1280,3)
56
+ 055 100 (720,1280,3)
57
+ 056 100 (720,1280,3)
58
+ 057 100 (720,1280,3)
59
+ 058 100 (720,1280,3)
60
+ 059 100 (720,1280,3)
61
+ 060 100 (720,1280,3)
62
+ 061 100 (720,1280,3)
63
+ 062 100 (720,1280,3)
64
+ 063 100 (720,1280,3)
65
+ 064 100 (720,1280,3)
66
+ 065 100 (720,1280,3)
67
+ 066 100 (720,1280,3)
68
+ 067 100 (720,1280,3)
69
+ 068 100 (720,1280,3)
70
+ 069 100 (720,1280,3)
71
+ 070 100 (720,1280,3)
72
+ 071 100 (720,1280,3)
73
+ 072 100 (720,1280,3)
74
+ 073 100 (720,1280,3)
75
+ 074 100 (720,1280,3)
76
+ 075 100 (720,1280,3)
77
+ 076 100 (720,1280,3)
78
+ 077 100 (720,1280,3)
79
+ 078 100 (720,1280,3)
80
+ 079 100 (720,1280,3)
81
+ 080 100 (720,1280,3)
82
+ 081 100 (720,1280,3)
83
+ 082 100 (720,1280,3)
84
+ 083 100 (720,1280,3)
85
+ 084 100 (720,1280,3)
86
+ 085 100 (720,1280,3)
87
+ 086 100 (720,1280,3)
88
+ 087 100 (720,1280,3)
89
+ 088 100 (720,1280,3)
90
+ 089 100 (720,1280,3)
91
+ 090 100 (720,1280,3)
92
+ 091 100 (720,1280,3)
93
+ 092 100 (720,1280,3)
94
+ 093 100 (720,1280,3)
95
+ 094 100 (720,1280,3)
96
+ 095 100 (720,1280,3)
97
+ 096 100 (720,1280,3)
98
+ 097 100 (720,1280,3)
99
+ 098 100 (720,1280,3)
100
+ 099 100 (720,1280,3)
101
+ 100 100 (720,1280,3)
102
+ 101 100 (720,1280,3)
103
+ 102 100 (720,1280,3)
104
+ 103 100 (720,1280,3)
105
+ 104 100 (720,1280,3)
106
+ 105 100 (720,1280,3)
107
+ 106 100 (720,1280,3)
108
+ 107 100 (720,1280,3)
109
+ 108 100 (720,1280,3)
110
+ 109 100 (720,1280,3)
111
+ 110 100 (720,1280,3)
112
+ 111 100 (720,1280,3)
113
+ 112 100 (720,1280,3)
114
+ 113 100 (720,1280,3)
115
+ 114 100 (720,1280,3)
116
+ 115 100 (720,1280,3)
117
+ 116 100 (720,1280,3)
118
+ 117 100 (720,1280,3)
119
+ 118 100 (720,1280,3)
120
+ 119 100 (720,1280,3)
121
+ 120 100 (720,1280,3)
122
+ 121 100 (720,1280,3)
123
+ 122 100 (720,1280,3)
124
+ 123 100 (720,1280,3)
125
+ 124 100 (720,1280,3)
126
+ 125 100 (720,1280,3)
127
+ 126 100 (720,1280,3)
128
+ 127 100 (720,1280,3)
129
+ 128 100 (720,1280,3)
130
+ 129 100 (720,1280,3)
131
+ 130 100 (720,1280,3)
132
+ 131 100 (720,1280,3)
133
+ 132 100 (720,1280,3)
134
+ 133 100 (720,1280,3)
135
+ 134 100 (720,1280,3)
136
+ 135 100 (720,1280,3)
137
+ 136 100 (720,1280,3)
138
+ 137 100 (720,1280,3)
139
+ 138 100 (720,1280,3)
140
+ 139 100 (720,1280,3)
141
+ 140 100 (720,1280,3)
142
+ 141 100 (720,1280,3)
143
+ 142 100 (720,1280,3)
144
+ 143 100 (720,1280,3)
145
+ 144 100 (720,1280,3)
146
+ 145 100 (720,1280,3)
147
+ 146 100 (720,1280,3)
148
+ 147 100 (720,1280,3)
149
+ 148 100 (720,1280,3)
150
+ 149 100 (720,1280,3)
151
+ 150 100 (720,1280,3)
152
+ 151 100 (720,1280,3)
153
+ 152 100 (720,1280,3)
154
+ 153 100 (720,1280,3)
155
+ 154 100 (720,1280,3)
156
+ 155 100 (720,1280,3)
157
+ 156 100 (720,1280,3)
158
+ 157 100 (720,1280,3)
159
+ 158 100 (720,1280,3)
160
+ 159 100 (720,1280,3)
161
+ 160 100 (720,1280,3)
162
+ 161 100 (720,1280,3)
163
+ 162 100 (720,1280,3)
164
+ 163 100 (720,1280,3)
165
+ 164 100 (720,1280,3)
166
+ 165 100 (720,1280,3)
167
+ 166 100 (720,1280,3)
168
+ 167 100 (720,1280,3)
169
+ 168 100 (720,1280,3)
170
+ 169 100 (720,1280,3)
171
+ 170 100 (720,1280,3)
172
+ 171 100 (720,1280,3)
173
+ 172 100 (720,1280,3)
174
+ 173 100 (720,1280,3)
175
+ 174 100 (720,1280,3)
176
+ 175 100 (720,1280,3)
177
+ 176 100 (720,1280,3)
178
+ 177 100 (720,1280,3)
179
+ 178 100 (720,1280,3)
180
+ 179 100 (720,1280,3)
181
+ 180 100 (720,1280,3)
182
+ 181 100 (720,1280,3)
183
+ 182 100 (720,1280,3)
184
+ 183 100 (720,1280,3)
185
+ 184 100 (720,1280,3)
186
+ 185 100 (720,1280,3)
187
+ 186 100 (720,1280,3)
188
+ 187 100 (720,1280,3)
189
+ 188 100 (720,1280,3)
190
+ 189 100 (720,1280,3)
191
+ 190 100 (720,1280,3)
192
+ 191 100 (720,1280,3)
193
+ 192 100 (720,1280,3)
194
+ 193 100 (720,1280,3)
195
+ 194 100 (720,1280,3)
196
+ 195 100 (720,1280,3)
197
+ 196 100 (720,1280,3)
198
+ 197 100 (720,1280,3)
199
+ 198 100 (720,1280,3)
200
+ 199 100 (720,1280,3)
201
+ 200 100 (720,1280,3)
202
+ 201 100 (720,1280,3)
203
+ 202 100 (720,1280,3)
204
+ 203 100 (720,1280,3)
205
+ 204 100 (720,1280,3)
206
+ 205 100 (720,1280,3)
207
+ 206 100 (720,1280,3)
208
+ 207 100 (720,1280,3)
209
+ 208 100 (720,1280,3)
210
+ 209 100 (720,1280,3)
211
+ 210 100 (720,1280,3)
212
+ 211 100 (720,1280,3)
213
+ 212 100 (720,1280,3)
214
+ 213 100 (720,1280,3)
215
+ 214 100 (720,1280,3)
216
+ 215 100 (720,1280,3)
217
+ 216 100 (720,1280,3)
218
+ 217 100 (720,1280,3)
219
+ 218 100 (720,1280,3)
220
+ 219 100 (720,1280,3)
221
+ 220 100 (720,1280,3)
222
+ 221 100 (720,1280,3)
223
+ 222 100 (720,1280,3)
224
+ 223 100 (720,1280,3)
225
+ 224 100 (720,1280,3)
226
+ 225 100 (720,1280,3)
227
+ 226 100 (720,1280,3)
228
+ 227 100 (720,1280,3)
229
+ 228 100 (720,1280,3)
230
+ 229 100 (720,1280,3)
231
+ 230 100 (720,1280,3)
232
+ 231 100 (720,1280,3)
233
+ 232 100 (720,1280,3)
234
+ 233 100 (720,1280,3)
235
+ 234 100 (720,1280,3)
236
+ 235 100 (720,1280,3)
237
+ 236 100 (720,1280,3)
238
+ 237 100 (720,1280,3)
239
+ 238 100 (720,1280,3)
240
+ 239 100 (720,1280,3)
241
+ 240 100 (720,1280,3)
242
+ 241 100 (720,1280,3)
243
+ 242 100 (720,1280,3)
244
+ 243 100 (720,1280,3)
245
+ 244 100 (720,1280,3)
246
+ 245 100 (720,1280,3)
247
+ 246 100 (720,1280,3)
248
+ 247 100 (720,1280,3)
249
+ 248 100 (720,1280,3)
250
+ 249 100 (720,1280,3)
251
+ 250 100 (720,1280,3)
252
+ 251 100 (720,1280,3)
253
+ 252 100 (720,1280,3)
254
+ 253 100 (720,1280,3)
255
+ 254 100 (720,1280,3)
256
+ 255 100 (720,1280,3)
257
+ 256 100 (720,1280,3)
258
+ 257 100 (720,1280,3)
259
+ 258 100 (720,1280,3)
260
+ 259 100 (720,1280,3)
261
+ 260 100 (720,1280,3)
262
+ 261 100 (720,1280,3)
263
+ 262 100 (720,1280,3)
264
+ 263 100 (720,1280,3)
265
+ 264 100 (720,1280,3)
266
+ 265 100 (720,1280,3)
267
+ 266 100 (720,1280,3)
268
+ 267 100 (720,1280,3)
269
+ 268 100 (720,1280,3)
270
+ 269 100 (720,1280,3)
basicsr/data/meta_info/meta_info_REDSofficial4_test_GT.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ 240 100 (720,1280,3)
2
+ 241 100 (720,1280,3)
3
+ 246 100 (720,1280,3)
4
+ 257 100 (720,1280,3)
basicsr/data/meta_info/meta_info_REDSval_official_test_GT.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 240 100 (720,1280,3)
2
+ 241 100 (720,1280,3)
3
+ 242 100 (720,1280,3)
4
+ 243 100 (720,1280,3)
5
+ 244 100 (720,1280,3)
6
+ 245 100 (720,1280,3)
7
+ 246 100 (720,1280,3)
8
+ 247 100 (720,1280,3)
9
+ 248 100 (720,1280,3)
10
+ 249 100 (720,1280,3)
11
+ 250 100 (720,1280,3)
12
+ 251 100 (720,1280,3)
13
+ 252 100 (720,1280,3)
14
+ 253 100 (720,1280,3)
15
+ 254 100 (720,1280,3)
16
+ 255 100 (720,1280,3)
17
+ 256 100 (720,1280,3)
18
+ 257 100 (720,1280,3)
19
+ 258 100 (720,1280,3)
20
+ 259 100 (720,1280,3)
21
+ 260 100 (720,1280,3)
22
+ 261 100 (720,1280,3)
23
+ 262 100 (720,1280,3)
24
+ 263 100 (720,1280,3)
25
+ 264 100 (720,1280,3)
26
+ 265 100 (720,1280,3)
27
+ 266 100 (720,1280,3)
28
+ 267 100 (720,1280,3)
29
+ 268 100 (720,1280,3)
30
+ 269 100 (720,1280,3)
basicsr/data/meta_info/meta_info_Vimeo90K_test_GT.txt ADDED
The diff for this file is too large to render. See raw diff
 
basicsr/data/meta_info/meta_info_Vimeo90K_test_fast_GT.txt ADDED
@@ -0,0 +1,1225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 00001/0625 7 (256,448,3)
2
+ 00001/0632 7 (256,448,3)
3
+ 00001/0807 7 (256,448,3)
4
+ 00001/0832 7 (256,448,3)
5
+ 00001/0834 7 (256,448,3)
6
+ 00001/0836 7 (256,448,3)
7
+ 00002/0004 7 (256,448,3)
8
+ 00002/0112 7 (256,448,3)
9
+ 00002/0116 7 (256,448,3)
10
+ 00002/0123 7 (256,448,3)
11
+ 00002/0455 7 (256,448,3)
12
+ 00002/0602 7 (256,448,3)
13
+ 00002/0976 7 (256,448,3)
14
+ 00002/0980 7 (256,448,3)
15
+ 00002/0983 7 (256,448,3)
16
+ 00002/1000 7 (256,448,3)
17
+ 00003/0022 7 (256,448,3)
18
+ 00003/0031 7 (256,448,3)
19
+ 00003/0035 7 (256,448,3)
20
+ 00003/0041 7 (256,448,3)
21
+ 00003/0073 7 (256,448,3)
22
+ 00003/0107 7 (256,448,3)
23
+ 00003/0111 7 (256,448,3)
24
+ 00003/0114 7 (256,448,3)
25
+ 00003/0117 7 (256,448,3)
26
+ 00003/0121 7 (256,448,3)
27
+ 00003/0499 7 (256,448,3)
28
+ 00003/0501 7 (256,448,3)
29
+ 00003/0507 7 (256,448,3)
30
+ 00003/0510 7 (256,448,3)
31
+ 00003/0517 7 (256,448,3)
32
+ 00003/0522 7 (256,448,3)
33
+ 00003/0531 7 (256,448,3)
34
+ 00003/0533 7 (256,448,3)
35
+ 00003/0534 7 (256,448,3)
36
+ 00003/0682 7 (256,448,3)
37
+ 00003/0687 7 (256,448,3)
38
+ 00003/0715 7 (256,448,3)
39
+ 00003/0742 7 (256,448,3)
40
+ 00003/0751 7 (256,448,3)
41
+ 00003/0984 7 (256,448,3)
42
+ 00004/0042 7 (256,448,3)
43
+ 00004/0165 7 (256,448,3)
44
+ 00004/0321 7 (256,448,3)
45
+ 00004/0569 7 (256,448,3)
46
+ 00004/0572 7 (256,448,3)
47
+ 00004/0619 7 (256,448,3)
48
+ 00004/0776 7 (256,448,3)
49
+ 00004/0780 7 (256,448,3)
50
+ 00004/0825 7 (256,448,3)
51
+ 00004/0832 7 (256,448,3)
52
+ 00004/0853 7 (256,448,3)
53
+ 00004/0876 7 (256,448,3)
54
+ 00004/0888 7 (256,448,3)
55
+ 00005/0015 7 (256,448,3)
56
+ 00005/0021 7 (256,448,3)
57
+ 00005/0022 7 (256,448,3)
58
+ 00005/0024 7 (256,448,3)
59
+ 00005/0026 7 (256,448,3)
60
+ 00005/0394 7 (256,448,3)
61
+ 00005/0403 7 (256,448,3)
62
+ 00005/0531 7 (256,448,3)
63
+ 00005/0546 7 (256,448,3)
64
+ 00005/0554 7 (256,448,3)
65
+ 00005/0694 7 (256,448,3)
66
+ 00005/0700 7 (256,448,3)
67
+ 00005/0740 7 (256,448,3)
68
+ 00005/0826 7 (256,448,3)
69
+ 00005/0832 7 (256,448,3)
70
+ 00005/0834 7 (256,448,3)
71
+ 00005/0943 7 (256,448,3)
72
+ 00006/0184 7 (256,448,3)
73
+ 00006/0205 7 (256,448,3)
74
+ 00006/0206 7 (256,448,3)
75
+ 00006/0211 7 (256,448,3)
76
+ 00006/0271 7 (256,448,3)
77
+ 00006/0273 7 (256,448,3)
78
+ 00006/0277 7 (256,448,3)
79
+ 00006/0283 7 (256,448,3)
80
+ 00006/0287 7 (256,448,3)
81
+ 00006/0298 7 (256,448,3)
82
+ 00006/0310 7 (256,448,3)
83
+ 00006/0356 7 (256,448,3)
84
+ 00006/0357 7 (256,448,3)
85
+ 00006/0544 7 (256,448,3)
86
+ 00006/0565 7 (256,448,3)
87
+ 00006/0569 7 (256,448,3)
88
+ 00006/0573 7 (256,448,3)
89
+ 00006/0592 7 (256,448,3)
90
+ 00006/0613 7 (256,448,3)
91
+ 00006/0633 7 (256,448,3)
92
+ 00006/0637 7 (256,448,3)
93
+ 00006/0646 7 (256,448,3)
94
+ 00006/0649 7 (256,448,3)
95
+ 00006/0655 7 (256,448,3)
96
+ 00006/0658 7 (256,448,3)
97
+ 00006/0662 7 (256,448,3)
98
+ 00006/0666 7 (256,448,3)
99
+ 00006/0673 7 (256,448,3)
100
+ 00007/0248 7 (256,448,3)
101
+ 00007/0253 7 (256,448,3)
102
+ 00007/0430 7 (256,448,3)
103
+ 00007/0434 7 (256,448,3)
104
+ 00007/0436 7 (256,448,3)
105
+ 00007/0452 7 (256,448,3)
106
+ 00007/0464 7 (256,448,3)
107
+ 00007/0470 7 (256,448,3)
108
+ 00007/0472 7 (256,448,3)
109
+ 00007/0483 7 (256,448,3)
110
+ 00007/0484 7 (256,448,3)
111
+ 00007/0493 7 (256,448,3)
112
+ 00007/0508 7 (256,448,3)
113
+ 00007/0514 7 (256,448,3)
114
+ 00007/0697 7 (256,448,3)
115
+ 00007/0698 7 (256,448,3)
116
+ 00007/0744 7 (256,448,3)
117
+ 00007/0775 7 (256,448,3)
118
+ 00007/0786 7 (256,448,3)
119
+ 00007/0790 7 (256,448,3)
120
+ 00007/0800 7 (256,448,3)
121
+ 00007/0833 7 (256,448,3)
122
+ 00007/0867 7 (256,448,3)
123
+ 00007/0879 7 (256,448,3)
124
+ 00007/0899 7 (256,448,3)
125
+ 00008/0251 7 (256,448,3)
126
+ 00008/0322 7 (256,448,3)
127
+ 00008/0971 7 (256,448,3)
128
+ 00008/0976 7 (256,448,3)
129
+ 00009/0016 7 (256,448,3)
130
+ 00009/0036 7 (256,448,3)
131
+ 00009/0037 7 (256,448,3)
132
+ 00009/0609 7 (256,448,3)
133
+ 00009/0812 7 (256,448,3)
134
+ 00009/0821 7 (256,448,3)
135
+ 00009/0947 7 (256,448,3)
136
+ 00009/0952 7 (256,448,3)
137
+ 00009/0955 7 (256,448,3)
138
+ 00009/0970 7 (256,448,3)
139
+ 00010/0072 7 (256,448,3)
140
+ 00010/0074 7 (256,448,3)
141
+ 00010/0079 7 (256,448,3)
142
+ 00010/0085 7 (256,448,3)
143
+ 00010/0139 7 (256,448,3)
144
+ 00010/0140 7 (256,448,3)
145
+ 00010/0183 7 (256,448,3)
146
+ 00010/0200 7 (256,448,3)
147
+ 00010/0223 7 (256,448,3)
148
+ 00010/0305 7 (256,448,3)
149
+ 00010/0323 7 (256,448,3)
150
+ 00010/0338 7 (256,448,3)
151
+ 00010/0342 7 (256,448,3)
152
+ 00010/0350 7 (256,448,3)
153
+ 00010/0356 7 (256,448,3)
154
+ 00010/0362 7 (256,448,3)
155
+ 00010/0366 7 (256,448,3)
156
+ 00010/0375 7 (256,448,3)
157
+ 00010/0404 7 (256,448,3)
158
+ 00010/0407 7 (256,448,3)
159
+ 00010/0414 7 (256,448,3)
160
+ 00010/0418 7 (256,448,3)
161
+ 00010/0429 7 (256,448,3)
162
+ 00010/0557 7 (256,448,3)
163
+ 00010/0564 7 (256,448,3)
164
+ 00010/0733 7 (256,448,3)
165
+ 00010/0935 7 (256,448,3)
166
+ 00010/0939 7 (256,448,3)
167
+ 00010/0943 7 (256,448,3)
168
+ 00011/0242 7 (256,448,3)
169
+ 00011/0259 7 (256,448,3)
170
+ 00011/0263 7 (256,448,3)
171
+ 00011/0266 7 (256,448,3)
172
+ 00011/0278 7 (256,448,3)
173
+ 00011/0890 7 (256,448,3)
174
+ 00011/0894 7 (256,448,3)
175
+ 00011/0903 7 (256,448,3)
176
+ 00011/0906 7 (256,448,3)
177
+ 00011/0913 7 (256,448,3)
178
+ 00012/0011 7 (256,448,3)
179
+ 00012/0014 7 (256,448,3)
180
+ 00012/0126 7 (256,448,3)
181
+ 00012/0127 7 (256,448,3)
182
+ 00012/0526 7 (256,448,3)
183
+ 00012/0551 7 (256,448,3)
184
+ 00012/0896 7 (256,448,3)
185
+ 00012/0910 7 (256,448,3)
186
+ 00012/0915 7 (256,448,3)
187
+ 00013/0167 7 (256,448,3)
188
+ 00013/0794 7 (256,448,3)
189
+ 00013/0807 7 (256,448,3)
190
+ 00013/0846 7 (256,448,3)
191
+ 00013/0882 7 (256,448,3)
192
+ 00013/0889 7 (256,448,3)
193
+ 00013/0910 7 (256,448,3)
194
+ 00013/0913 7 (256,448,3)
195
+ 00013/0924 7 (256,448,3)
196
+ 00013/0931 7 (256,448,3)
197
+ 00013/0944 7 (256,448,3)
198
+ 00013/0955 7 (256,448,3)
199
+ 00013/0962 7 (256,448,3)
200
+ 00013/0969 7 (256,448,3)
201
+ 00014/0012 7 (256,448,3)
202
+ 00014/0025 7 (256,448,3)
203
+ 00014/0473 7 (256,448,3)
204
+ 00014/0499 7 (256,448,3)
205
+ 00014/0524 7 (256,448,3)
206
+ 00014/0739 7 (256,448,3)
207
+ 00014/0753 7 (256,448,3)
208
+ 00014/0771 7 (256,448,3)
209
+ 00014/0832 7 (256,448,3)
210
+ 00014/0836 7 (256,448,3)
211
+ 00014/0838 7 (256,448,3)
212
+ 00014/0839 7 (256,448,3)
213
+ 00014/0843 7 (256,448,3)
214
+ 00014/0846 7 (256,448,3)
215
+ 00014/0849 7 (256,448,3)
216
+ 00014/0859 7 (256,448,3)
217
+ 00014/0880 7 (256,448,3)
218
+ 00014/0906 7 (256,448,3)
219
+ 00015/0030 7 (256,448,3)
220
+ 00015/0067 7 (256,448,3)
221
+ 00015/0084 7 (256,448,3)
222
+ 00015/0190 7 (256,448,3)
223
+ 00015/0575 7 (256,448,3)
224
+ 00015/0784 7 (256,448,3)
225
+ 00015/0855 7 (256,448,3)
226
+ 00015/0904 7 (256,448,3)
227
+ 00015/0914 7 (256,448,3)
228
+ 00015/0936 7 (256,448,3)
229
+ 00015/0939 7 (256,448,3)
230
+ 00015/0943 7 (256,448,3)
231
+ 00015/0957 7 (256,448,3)
232
+ 00016/0131 7 (256,448,3)
233
+ 00016/0173 7 (256,448,3)
234
+ 00016/0320 7 (256,448,3)
235
+ 00016/0328 7 (256,448,3)
236
+ 00016/0334 7 (256,448,3)
237
+ 00016/0338 7 (256,448,3)
238
+ 00016/0339 7 (256,448,3)
239
+ 00016/0345 7 (256,448,3)
240
+ 00016/0365 7 (256,448,3)
241
+ 00016/0584 7 (256,448,3)
242
+ 00016/0634 7 (256,448,3)
243
+ 00017/0342 7 (256,448,3)
244
+ 00017/0346 7 (256,448,3)
245
+ 00017/0350 7 (256,448,3)
246
+ 00017/0766 7 (256,448,3)
247
+ 00017/0786 7 (256,448,3)
248
+ 00017/0911 7 (256,448,3)
249
+ 00017/0914 7 (256,448,3)
250
+ 00018/0217 7 (256,448,3)
251
+ 00018/0258 7 (256,448,3)
252
+ 00018/0307 7 (256,448,3)
253
+ 00018/0480 7 (256,448,3)
254
+ 00018/0491 7 (256,448,3)
255
+ 00018/0994 7 (256,448,3)
256
+ 00018/0995 7 (256,448,3)
257
+ 00018/0997 7 (256,448,3)
258
+ 00018/1000 7 (256,448,3)
259
+ 00019/0007 7 (256,448,3)
260
+ 00019/0016 7 (256,448,3)
261
+ 00019/0026 7 (256,448,3)
262
+ 00019/0030 7 (256,448,3)
263
+ 00019/0086 7 (256,448,3)
264
+ 00019/0089 7 (256,448,3)
265
+ 00019/0111 7 (256,448,3)
266
+ 00019/0285 7 (256,448,3)
267
+ 00019/0415 7 (256,448,3)
268
+ 00019/0434 7 (256,448,3)
269
+ 00019/0437 7 (256,448,3)
270
+ 00019/0568 7 (256,448,3)
271
+ 00019/0570 7 (256,448,3)
272
+ 00019/0591 7 (256,448,3)
273
+ 00019/0596 7 (256,448,3)
274
+ 00019/0603 7 (256,448,3)
275
+ 00019/0607 7 (256,448,3)
276
+ 00019/0637 7 (256,448,3)
277
+ 00019/0644 7 (256,448,3)
278
+ 00019/0647 7 (256,448,3)
279
+ 00019/0787 7 (256,448,3)
280
+ 00019/0993 7 (256,448,3)
281
+ 00019/0998 7 (256,448,3)
282
+ 00021/0232 7 (256,448,3)
283
+ 00021/0255 7 (256,448,3)
284
+ 00021/0646 7 (256,448,3)
285
+ 00021/0653 7 (256,448,3)
286
+ 00021/0657 7 (256,448,3)
287
+ 00021/0668 7 (256,448,3)
288
+ 00021/0672 7 (256,448,3)
289
+ 00021/0725 7 (256,448,3)
290
+ 00021/0750 7 (256,448,3)
291
+ 00021/0764 7 (256,448,3)
292
+ 00021/0821 7 (256,448,3)
293
+ 00022/0192 7 (256,448,3)
294
+ 00022/0391 7 (256,448,3)
295
+ 00022/0514 7 (256,448,3)
296
+ 00022/0567 7 (256,448,3)
297
+ 00022/0674 7 (256,448,3)
298
+ 00022/0686 7 (256,448,3)
299
+ 00022/0700 7 (256,448,3)
300
+ 00023/0020 7 (256,448,3)
301
+ 00023/0024 7 (256,448,3)
302
+ 00023/0025 7 (256,448,3)
303
+ 00023/0042 7 (256,448,3)
304
+ 00023/0050 7 (256,448,3)
305
+ 00023/0094 7 (256,448,3)
306
+ 00023/0107 7 (256,448,3)
307
+ 00023/0635 7 (256,448,3)
308
+ 00023/0698 7 (256,448,3)
309
+ 00023/0774 7 (256,448,3)
310
+ 00023/0795 7 (256,448,3)
311
+ 00023/0821 7 (256,448,3)
312
+ 00023/0839 7 (256,448,3)
313
+ 00023/0846 7 (256,448,3)
314
+ 00023/0869 7 (256,448,3)
315
+ 00023/0879 7 (256,448,3)
316
+ 00023/0887 7 (256,448,3)
317
+ 00023/0899 7 (256,448,3)
318
+ 00023/0910 7 (256,448,3)
319
+ 00023/0920 7 (256,448,3)
320
+ 00023/0929 7 (256,448,3)
321
+ 00023/0941 7 (256,448,3)
322
+ 00023/0942 7 (256,448,3)
323
+ 00023/0952 7 (256,448,3)
324
+ 00024/0066 7 (256,448,3)
325
+ 00024/0072 7 (256,448,3)
326
+ 00024/0080 7 (256,448,3)
327
+ 00024/0093 7 (256,448,3)
328
+ 00024/0107 7 (256,448,3)
329
+ 00024/0262 7 (256,448,3)
330
+ 00024/0283 7 (256,448,3)
331
+ 00024/0294 7 (256,448,3)
332
+ 00024/0296 7 (256,448,3)
333
+ 00024/0304 7 (256,448,3)
334
+ 00024/0315 7 (256,448,3)
335
+ 00024/0322 7 (256,448,3)
336
+ 00024/0648 7 (256,448,3)
337
+ 00024/0738 7 (256,448,3)
338
+ 00024/0743 7 (256,448,3)
339
+ 00025/0542 7 (256,448,3)
340
+ 00025/0769 7 (256,448,3)
341
+ 00025/0984 7 (256,448,3)
342
+ 00025/0985 7 (256,448,3)
343
+ 00025/0989 7 (256,448,3)
344
+ 00025/0991 7 (256,448,3)
345
+ 00026/0009 7 (256,448,3)
346
+ 00026/0013 7 (256,448,3)
347
+ 00026/0020 7 (256,448,3)
348
+ 00026/0021 7 (256,448,3)
349
+ 00026/0025 7 (256,448,3)
350
+ 00026/0135 7 (256,448,3)
351
+ 00026/0200 7 (256,448,3)
352
+ 00026/0297 7 (256,448,3)
353
+ 00026/0306 7 (256,448,3)
354
+ 00026/0444 7 (256,448,3)
355
+ 00026/0450 7 (256,448,3)
356
+ 00026/0453 7 (256,448,3)
357
+ 00026/0464 7 (256,448,3)
358
+ 00026/0486 7 (256,448,3)
359
+ 00026/0773 7 (256,448,3)
360
+ 00026/0785 7 (256,448,3)
361
+ 00026/0836 7 (256,448,3)
362
+ 00026/0838 7 (256,448,3)
363
+ 00026/0848 7 (256,448,3)
364
+ 00026/0885 7 (256,448,3)
365
+ 00026/0893 7 (256,448,3)
366
+ 00026/0939 7 (256,448,3)
367
+ 00026/0942 7 (256,448,3)
368
+ 00027/0092 7 (256,448,3)
369
+ 00027/0112 7 (256,448,3)
370
+ 00027/0115 7 (256,448,3)
371
+ 00027/0143 7 (256,448,3)
372
+ 00027/0175 7 (256,448,3)
373
+ 00027/0179 7 (256,448,3)
374
+ 00027/0183 7 (256,448,3)
375
+ 00027/0197 7 (256,448,3)
376
+ 00027/0199 7 (256,448,3)
377
+ 00027/0300 7 (256,448,3)
378
+ 00028/0015 7 (256,448,3)
379
+ 00028/0032 7 (256,448,3)
380
+ 00028/0048 7 (256,448,3)
381
+ 00028/0068 7 (256,448,3)
382
+ 00028/0219 7 (256,448,3)
383
+ 00028/0606 7 (256,448,3)
384
+ 00028/0626 7 (256,448,3)
385
+ 00028/0748 7 (256,448,3)
386
+ 00028/0764 7 (256,448,3)
387
+ 00028/0772 7 (256,448,3)
388
+ 00028/0780 7 (256,448,3)
389
+ 00028/0926 7 (256,448,3)
390
+ 00028/0947 7 (256,448,3)
391
+ 00028/0962 7 (256,448,3)
392
+ 00029/0085 7 (256,448,3)
393
+ 00029/0281 7 (256,448,3)
394
+ 00029/0284 7 (256,448,3)
395
+ 00029/0288 7 (256,448,3)
396
+ 00029/0294 7 (256,448,3)
397
+ 00029/0364 7 (256,448,3)
398
+ 00029/0369 7 (256,448,3)
399
+ 00029/0421 7 (256,448,3)
400
+ 00029/0425 7 (256,448,3)
401
+ 00029/0550 7 (256,448,3)
402
+ 00030/0014 7 (256,448,3)
403
+ 00030/0101 7 (256,448,3)
404
+ 00030/0143 7 (256,448,3)
405
+ 00030/0351 7 (256,448,3)
406
+ 00030/0356 7 (256,448,3)
407
+ 00030/0371 7 (256,448,3)
408
+ 00030/0484 7 (256,448,3)
409
+ 00030/0492 7 (256,448,3)
410
+ 00030/0503 7 (256,448,3)
411
+ 00030/0682 7 (256,448,3)
412
+ 00030/0696 7 (256,448,3)
413
+ 00030/0735 7 (256,448,3)
414
+ 00030/0737 7 (256,448,3)
415
+ 00030/0868 7 (256,448,3)
416
+ 00031/0161 7 (256,448,3)
417
+ 00031/0180 7 (256,448,3)
418
+ 00031/0194 7 (256,448,3)
419
+ 00031/0253 7 (256,448,3)
420
+ 00031/0293 7 (256,448,3)
421
+ 00031/0466 7 (256,448,3)
422
+ 00031/0477 7 (256,448,3)
423
+ 00031/0549 7 (256,448,3)
424
+ 00031/0600 7 (256,448,3)
425
+ 00031/0617 7 (256,448,3)
426
+ 00031/0649 7 (256,448,3)
427
+ 00032/0015 7 (256,448,3)
428
+ 00032/0020 7 (256,448,3)
429
+ 00032/0023 7 (256,448,3)
430
+ 00032/0048 7 (256,448,3)
431
+ 00032/0056 7 (256,448,3)
432
+ 00032/0872 7 (256,448,3)
433
+ 00033/0069 7 (256,448,3)
434
+ 00033/0073 7 (256,448,3)
435
+ 00033/0078 7 (256,448,3)
436
+ 00033/0079 7 (256,448,3)
437
+ 00033/0086 7 (256,448,3)
438
+ 00033/0088 7 (256,448,3)
439
+ 00033/0091 7 (256,448,3)
440
+ 00033/0096 7 (256,448,3)
441
+ 00033/0607 7 (256,448,3)
442
+ 00033/0613 7 (256,448,3)
443
+ 00033/0616 7 (256,448,3)
444
+ 00033/0619 7 (256,448,3)
445
+ 00033/0626 7 (256,448,3)
446
+ 00033/0628 7 (256,448,3)
447
+ 00033/0637 7 (256,448,3)
448
+ 00033/0686 7 (256,448,3)
449
+ 00033/0842 7 (256,448,3)
450
+ 00034/0261 7 (256,448,3)
451
+ 00034/0265 7 (256,448,3)
452
+ 00034/0269 7 (256,448,3)
453
+ 00034/0275 7 (256,448,3)
454
+ 00034/0286 7 (256,448,3)
455
+ 00034/0294 7 (256,448,3)
456
+ 00034/0431 7 (256,448,3)
457
+ 00034/0577 7 (256,448,3)
458
+ 00034/0685 7 (256,448,3)
459
+ 00034/0687 7 (256,448,3)
460
+ 00034/0703 7 (256,448,3)
461
+ 00034/0715 7 (256,448,3)
462
+ 00034/0935 7 (256,448,3)
463
+ 00034/0943 7 (256,448,3)
464
+ 00034/0963 7 (256,448,3)
465
+ 00034/0979 7 (256,448,3)
466
+ 00034/0990 7 (256,448,3)
467
+ 00035/0129 7 (256,448,3)
468
+ 00035/0153 7 (256,448,3)
469
+ 00035/0156 7 (256,448,3)
470
+ 00035/0474 7 (256,448,3)
471
+ 00035/0507 7 (256,448,3)
472
+ 00035/0532 7 (256,448,3)
473
+ 00035/0560 7 (256,448,3)
474
+ 00035/0572 7 (256,448,3)
475
+ 00035/0587 7 (256,448,3)
476
+ 00035/0588 7 (256,448,3)
477
+ 00035/0640 7 (256,448,3)
478
+ 00035/0654 7 (256,448,3)
479
+ 00035/0655 7 (256,448,3)
480
+ 00035/0737 7 (256,448,3)
481
+ 00035/0843 7 (256,448,3)
482
+ 00035/0932 7 (256,448,3)
483
+ 00035/0957 7 (256,448,3)
484
+ 00036/0029 7 (256,448,3)
485
+ 00036/0266 7 (256,448,3)
486
+ 00036/0276 7 (256,448,3)
487
+ 00036/0310 7 (256,448,3)
488
+ 00036/0314 7 (256,448,3)
489
+ 00036/0320 7 (256,448,3)
490
+ 00036/0333 7 (256,448,3)
491
+ 00036/0348 7 (256,448,3)
492
+ 00036/0357 7 (256,448,3)
493
+ 00036/0360 7 (256,448,3)
494
+ 00036/0368 7 (256,448,3)
495
+ 00036/0371 7 (256,448,3)
496
+ 00036/0378 7 (256,448,3)
497
+ 00036/0391 7 (256,448,3)
498
+ 00036/0440 7 (256,448,3)
499
+ 00036/0731 7 (256,448,3)
500
+ 00036/0733 7 (256,448,3)
501
+ 00036/0741 7 (256,448,3)
502
+ 00036/0743 7 (256,448,3)
503
+ 00036/0927 7 (256,448,3)
504
+ 00036/0931 7 (256,448,3)
505
+ 00036/0933 7 (256,448,3)
506
+ 00036/0938 7 (256,448,3)
507
+ 00036/0944 7 (256,448,3)
508
+ 00036/0946 7 (256,448,3)
509
+ 00036/0951 7 (256,448,3)
510
+ 00036/0953 7 (256,448,3)
511
+ 00036/0963 7 (256,448,3)
512
+ 00036/0964 7 (256,448,3)
513
+ 00036/0981 7 (256,448,3)
514
+ 00036/0991 7 (256,448,3)
515
+ 00037/0072 7 (256,448,3)
516
+ 00037/0079 7 (256,448,3)
517
+ 00037/0132 7 (256,448,3)
518
+ 00037/0135 7 (256,448,3)
519
+ 00037/0137 7 (256,448,3)
520
+ 00037/0141 7 (256,448,3)
521
+ 00037/0229 7 (256,448,3)
522
+ 00037/0234 7 (256,448,3)
523
+ 00037/0239 7 (256,448,3)
524
+ 00037/0242 7 (256,448,3)
525
+ 00037/0254 7 (256,448,3)
526
+ 00037/0269 7 (256,448,3)
527
+ 00037/0276 7 (256,448,3)
528
+ 00037/0279 7 (256,448,3)
529
+ 00037/0286 7 (256,448,3)
530
+ 00037/0345 7 (256,448,3)
531
+ 00037/0449 7 (256,448,3)
532
+ 00037/0450 7 (256,448,3)
533
+ 00037/0820 7 (256,448,3)
534
+ 00037/0824 7 (256,448,3)
535
+ 00037/0859 7 (256,448,3)
536
+ 00037/0899 7 (256,448,3)
537
+ 00037/0906 7 (256,448,3)
538
+ 00038/0535 7 (256,448,3)
539
+ 00038/0572 7 (256,448,3)
540
+ 00038/0675 7 (256,448,3)
541
+ 00038/0731 7 (256,448,3)
542
+ 00038/0732 7 (256,448,3)
543
+ 00038/0744 7 (256,448,3)
544
+ 00038/0755 7 (256,448,3)
545
+ 00039/0002 7 (256,448,3)
546
+ 00039/0013 7 (256,448,3)
547
+ 00039/0247 7 (256,448,3)
548
+ 00039/0489 7 (256,448,3)
549
+ 00039/0504 7 (256,448,3)
550
+ 00039/0558 7 (256,448,3)
551
+ 00039/0686 7 (256,448,3)
552
+ 00039/0727 7 (256,448,3)
553
+ 00039/0769 7 (256,448,3)
554
+ 00040/0081 7 (256,448,3)
555
+ 00040/0082 7 (256,448,3)
556
+ 00040/0402 7 (256,448,3)
557
+ 00040/0407 7 (256,448,3)
558
+ 00040/0408 7 (256,448,3)
559
+ 00040/0410 7 (256,448,3)
560
+ 00040/0411 7 (256,448,3)
561
+ 00040/0412 7 (256,448,3)
562
+ 00040/0413 7 (256,448,3)
563
+ 00040/0415 7 (256,448,3)
564
+ 00040/0421 7 (256,448,3)
565
+ 00040/0422 7 (256,448,3)
566
+ 00040/0426 7 (256,448,3)
567
+ 00040/0438 7 (256,448,3)
568
+ 00040/0439 7 (256,448,3)
569
+ 00040/0440 7 (256,448,3)
570
+ 00040/0443 7 (256,448,3)
571
+ 00040/0457 7 (256,448,3)
572
+ 00040/0459 7 (256,448,3)
573
+ 00040/0725 7 (256,448,3)
574
+ 00040/0727 7 (256,448,3)
575
+ 00040/0936 7 (256,448,3)
576
+ 00040/0959 7 (256,448,3)
577
+ 00040/0964 7 (256,448,3)
578
+ 00040/0968 7 (256,448,3)
579
+ 00040/0974 7 (256,448,3)
580
+ 00040/0978 7 (256,448,3)
581
+ 00040/0979 7 (256,448,3)
582
+ 00040/0989 7 (256,448,3)
583
+ 00040/0993 7 (256,448,3)
584
+ 00040/0994 7 (256,448,3)
585
+ 00040/0997 7 (256,448,3)
586
+ 00041/0001 7 (256,448,3)
587
+ 00041/0007 7 (256,448,3)
588
+ 00041/0019 7 (256,448,3)
589
+ 00041/0040 7 (256,448,3)
590
+ 00041/0350 7 (256,448,3)
591
+ 00041/0357 7 (256,448,3)
592
+ 00041/0393 7 (256,448,3)
593
+ 00041/0890 7 (256,448,3)
594
+ 00041/0909 7 (256,448,3)
595
+ 00041/0915 7 (256,448,3)
596
+ 00041/0933 7 (256,448,3)
597
+ 00042/0017 7 (256,448,3)
598
+ 00042/0332 7 (256,448,3)
599
+ 00042/0346 7 (256,448,3)
600
+ 00042/0350 7 (256,448,3)
601
+ 00042/0356 7 (256,448,3)
602
+ 00042/0382 7 (256,448,3)
603
+ 00042/0389 7 (256,448,3)
604
+ 00042/0539 7 (256,448,3)
605
+ 00042/0546 7 (256,448,3)
606
+ 00042/0550 7 (256,448,3)
607
+ 00042/0553 7 (256,448,3)
608
+ 00042/0555 7 (256,448,3)
609
+ 00042/0560 7 (256,448,3)
610
+ 00042/0570 7 (256,448,3)
611
+ 00043/0119 7 (256,448,3)
612
+ 00043/0122 7 (256,448,3)
613
+ 00043/0168 7 (256,448,3)
614
+ 00043/0274 7 (256,448,3)
615
+ 00043/0304 7 (256,448,3)
616
+ 00043/0731 7 (256,448,3)
617
+ 00043/0735 7 (256,448,3)
618
+ 00043/0739 7 (256,448,3)
619
+ 00043/0740 7 (256,448,3)
620
+ 00044/0212 7 (256,448,3)
621
+ 00044/0432 7 (256,448,3)
622
+ 00044/0934 7 (256,448,3)
623
+ 00044/0940 7 (256,448,3)
624
+ 00044/0987 7 (256,448,3)
625
+ 00045/0004 7 (256,448,3)
626
+ 00045/0009 7 (256,448,3)
627
+ 00045/0011 7 (256,448,3)
628
+ 00045/0019 7 (256,448,3)
629
+ 00045/0023 7 (256,448,3)
630
+ 00045/0289 7 (256,448,3)
631
+ 00045/0760 7 (256,448,3)
632
+ 00045/0779 7 (256,448,3)
633
+ 00045/0816 7 (256,448,3)
634
+ 00045/0820 7 (256,448,3)
635
+ 00046/0132 7 (256,448,3)
636
+ 00046/0350 7 (256,448,3)
637
+ 00046/0356 7 (256,448,3)
638
+ 00046/0357 7 (256,448,3)
639
+ 00046/0379 7 (256,448,3)
640
+ 00046/0410 7 (256,448,3)
641
+ 00046/0412 7 (256,448,3)
642
+ 00046/0481 7 (256,448,3)
643
+ 00046/0497 7 (256,448,3)
644
+ 00046/0510 7 (256,448,3)
645
+ 00046/0515 7 (256,448,3)
646
+ 00046/0529 7 (256,448,3)
647
+ 00046/0544 7 (256,448,3)
648
+ 00046/0545 7 (256,448,3)
649
+ 00046/0552 7 (256,448,3)
650
+ 00046/0559 7 (256,448,3)
651
+ 00046/0589 7 (256,448,3)
652
+ 00046/0642 7 (256,448,3)
653
+ 00046/0724 7 (256,448,3)
654
+ 00046/0758 7 (256,448,3)
655
+ 00046/0930 7 (256,448,3)
656
+ 00046/0953 7 (256,448,3)
657
+ 00047/0013 7 (256,448,3)
658
+ 00047/0014 7 (256,448,3)
659
+ 00047/0017 7 (256,448,3)
660
+ 00047/0076 7 (256,448,3)
661
+ 00047/0151 7 (256,448,3)
662
+ 00047/0797 7 (256,448,3)
663
+ 00048/0014 7 (256,448,3)
664
+ 00048/0021 7 (256,448,3)
665
+ 00048/0026 7 (256,448,3)
666
+ 00048/0030 7 (256,448,3)
667
+ 00048/0039 7 (256,448,3)
668
+ 00048/0045 7 (256,448,3)
669
+ 00048/0049 7 (256,448,3)
670
+ 00048/0145 7 (256,448,3)
671
+ 00048/0188 7 (256,448,3)
672
+ 00048/0302 7 (256,448,3)
673
+ 00048/0361 7 (256,448,3)
674
+ 00048/0664 7 (256,448,3)
675
+ 00048/0672 7 (256,448,3)
676
+ 00048/0681 7 (256,448,3)
677
+ 00048/0689 7 (256,448,3)
678
+ 00048/0690 7 (256,448,3)
679
+ 00048/0691 7 (256,448,3)
680
+ 00048/0711 7 (256,448,3)
681
+ 00049/0085 7 (256,448,3)
682
+ 00049/0810 7 (256,448,3)
683
+ 00049/0858 7 (256,448,3)
684
+ 00049/0865 7 (256,448,3)
685
+ 00049/0871 7 (256,448,3)
686
+ 00049/0903 7 (256,448,3)
687
+ 00049/0928 7 (256,448,3)
688
+ 00050/0092 7 (256,448,3)
689
+ 00050/0101 7 (256,448,3)
690
+ 00050/0108 7 (256,448,3)
691
+ 00050/0112 7 (256,448,3)
692
+ 00050/0120 7 (256,448,3)
693
+ 00050/0128 7 (256,448,3)
694
+ 00050/0383 7 (256,448,3)
695
+ 00050/0395 7 (256,448,3)
696
+ 00050/0405 7 (256,448,3)
697
+ 00050/0632 7 (256,448,3)
698
+ 00050/0648 7 (256,448,3)
699
+ 00050/0649 7 (256,448,3)
700
+ 00050/0659 7 (256,448,3)
701
+ 00050/0699 7 (256,448,3)
702
+ 00050/0708 7 (256,448,3)
703
+ 00050/0716 7 (256,448,3)
704
+ 00050/0758 7 (256,448,3)
705
+ 00050/0761 7 (256,448,3)
706
+ 00051/0572 7 (256,448,3)
707
+ 00052/0163 7 (256,448,3)
708
+ 00052/0242 7 (256,448,3)
709
+ 00052/0260 7 (256,448,3)
710
+ 00052/0322 7 (256,448,3)
711
+ 00052/0333 7 (256,448,3)
712
+ 00052/0806 7 (256,448,3)
713
+ 00052/0813 7 (256,448,3)
714
+ 00052/0821 7 (256,448,3)
715
+ 00052/0830 7 (256,448,3)
716
+ 00052/0914 7 (256,448,3)
717
+ 00052/0923 7 (256,448,3)
718
+ 00052/0959 7 (256,448,3)
719
+ 00053/0288 7 (256,448,3)
720
+ 00053/0290 7 (256,448,3)
721
+ 00053/0323 7 (256,448,3)
722
+ 00053/0337 7 (256,448,3)
723
+ 00053/0340 7 (256,448,3)
724
+ 00053/0437 7 (256,448,3)
725
+ 00053/0595 7 (256,448,3)
726
+ 00053/0739 7 (256,448,3)
727
+ 00053/0761 7 (256,448,3)
728
+ 00054/0014 7 (256,448,3)
729
+ 00054/0017 7 (256,448,3)
730
+ 00054/0178 7 (256,448,3)
731
+ 00054/0183 7 (256,448,3)
732
+ 00054/0196 7 (256,448,3)
733
+ 00054/0205 7 (256,448,3)
734
+ 00054/0214 7 (256,448,3)
735
+ 00054/0289 7 (256,448,3)
736
+ 00054/0453 7 (256,448,3)
737
+ 00054/0498 7 (256,448,3)
738
+ 00054/0502 7 (256,448,3)
739
+ 00054/0514 7 (256,448,3)
740
+ 00054/0773 7 (256,448,3)
741
+ 00055/0001 7 (256,448,3)
742
+ 00055/0115 7 (256,448,3)
743
+ 00055/0118 7 (256,448,3)
744
+ 00055/0171 7 (256,448,3)
745
+ 00055/0214 7 (256,448,3)
746
+ 00055/0354 7 (256,448,3)
747
+ 00055/0449 7 (256,448,3)
748
+ 00055/0473 7 (256,448,3)
749
+ 00055/0649 7 (256,448,3)
750
+ 00055/0800 7 (256,448,3)
751
+ 00055/0803 7 (256,448,3)
752
+ 00055/0990 7 (256,448,3)
753
+ 00056/0041 7 (256,448,3)
754
+ 00056/0120 7 (256,448,3)
755
+ 00056/0293 7 (256,448,3)
756
+ 00056/0357 7 (256,448,3)
757
+ 00056/0506 7 (256,448,3)
758
+ 00056/0561 7 (256,448,3)
759
+ 00056/0567 7 (256,448,3)
760
+ 00056/0575 7 (256,448,3)
761
+ 00057/0175 7 (256,448,3)
762
+ 00057/0495 7 (256,448,3)
763
+ 00057/0498 7 (256,448,3)
764
+ 00057/0506 7 (256,448,3)
765
+ 00057/0612 7 (256,448,3)
766
+ 00057/0620 7 (256,448,3)
767
+ 00057/0623 7 (256,448,3)
768
+ 00057/0635 7 (256,448,3)
769
+ 00057/0773 7 (256,448,3)
770
+ 00057/0778 7 (256,448,3)
771
+ 00057/0867 7 (256,448,3)
772
+ 00057/0976 7 (256,448,3)
773
+ 00057/0980 7 (256,448,3)
774
+ 00057/0985 7 (256,448,3)
775
+ 00057/0992 7 (256,448,3)
776
+ 00058/0009 7 (256,448,3)
777
+ 00058/0076 7 (256,448,3)
778
+ 00058/0078 7 (256,448,3)
779
+ 00058/0279 7 (256,448,3)
780
+ 00058/0283 7 (256,448,3)
781
+ 00058/0286 7 (256,448,3)
782
+ 00058/0350 7 (256,448,3)
783
+ 00058/0380 7 (256,448,3)
784
+ 00061/0132 7 (256,448,3)
785
+ 00061/0141 7 (256,448,3)
786
+ 00061/0156 7 (256,448,3)
787
+ 00061/0159 7 (256,448,3)
788
+ 00061/0168 7 (256,448,3)
789
+ 00061/0170 7 (256,448,3)
790
+ 00061/0186 7 (256,448,3)
791
+ 00061/0219 7 (256,448,3)
792
+ 00061/0227 7 (256,448,3)
793
+ 00061/0238 7 (256,448,3)
794
+ 00061/0256 7 (256,448,3)
795
+ 00061/0303 7 (256,448,3)
796
+ 00061/0312 7 (256,448,3)
797
+ 00061/0313 7 (256,448,3)
798
+ 00061/0325 7 (256,448,3)
799
+ 00061/0367 7 (256,448,3)
800
+ 00061/0369 7 (256,448,3)
801
+ 00061/0387 7 (256,448,3)
802
+ 00061/0396 7 (256,448,3)
803
+ 00061/0486 7 (256,448,3)
804
+ 00061/0895 7 (256,448,3)
805
+ 00061/0897 7 (256,448,3)
806
+ 00062/0846 7 (256,448,3)
807
+ 00063/0156 7 (256,448,3)
808
+ 00063/0184 7 (256,448,3)
809
+ 00063/0191 7 (256,448,3)
810
+ 00063/0334 7 (256,448,3)
811
+ 00063/0350 7 (256,448,3)
812
+ 00063/0499 7 (256,448,3)
813
+ 00063/0878 7 (256,448,3)
814
+ 00064/0004 7 (256,448,3)
815
+ 00064/0264 7 (256,448,3)
816
+ 00064/0735 7 (256,448,3)
817
+ 00064/0738 7 (256,448,3)
818
+ 00065/0105 7 (256,448,3)
819
+ 00065/0169 7 (256,448,3)
820
+ 00065/0305 7 (256,448,3)
821
+ 00065/0324 7 (256,448,3)
822
+ 00065/0353 7 (256,448,3)
823
+ 00065/0520 7 (256,448,3)
824
+ 00065/0533 7 (256,448,3)
825
+ 00065/0545 7 (256,448,3)
826
+ 00065/0551 7 (256,448,3)
827
+ 00065/0568 7 (256,448,3)
828
+ 00065/0603 7 (256,448,3)
829
+ 00065/0884 7 (256,448,3)
830
+ 00065/0988 7 (256,448,3)
831
+ 00066/0002 7 (256,448,3)
832
+ 00066/0011 7 (256,448,3)
833
+ 00066/0031 7 (256,448,3)
834
+ 00066/0037 7 (256,448,3)
835
+ 00066/0136 7 (256,448,3)
836
+ 00066/0137 7 (256,448,3)
837
+ 00066/0150 7 (256,448,3)
838
+ 00066/0166 7 (256,448,3)
839
+ 00066/0178 7 (256,448,3)
840
+ 00066/0357 7 (256,448,3)
841
+ 00066/0428 7 (256,448,3)
842
+ 00066/0483 7 (256,448,3)
843
+ 00066/0600 7 (256,448,3)
844
+ 00066/0863 7 (256,448,3)
845
+ 00066/0873 7 (256,448,3)
846
+ 00066/0875 7 (256,448,3)
847
+ 00066/0899 7 (256,448,3)
848
+ 00067/0020 7 (256,448,3)
849
+ 00067/0025 7 (256,448,3)
850
+ 00067/0132 7 (256,448,3)
851
+ 00067/0492 7 (256,448,3)
852
+ 00067/0726 7 (256,448,3)
853
+ 00067/0734 7 (256,448,3)
854
+ 00067/0744 7 (256,448,3)
855
+ 00067/0754 7 (256,448,3)
856
+ 00067/0779 7 (256,448,3)
857
+ 00068/0078 7 (256,448,3)
858
+ 00068/0083 7 (256,448,3)
859
+ 00068/0113 7 (256,448,3)
860
+ 00068/0117 7 (256,448,3)
861
+ 00068/0121 7 (256,448,3)
862
+ 00068/0206 7 (256,448,3)
863
+ 00068/0261 7 (256,448,3)
864
+ 00068/0321 7 (256,448,3)
865
+ 00068/0354 7 (256,448,3)
866
+ 00068/0380 7 (256,448,3)
867
+ 00068/0419 7 (256,448,3)
868
+ 00068/0547 7 (256,448,3)
869
+ 00068/0561 7 (256,448,3)
870
+ 00068/0565 7 (256,448,3)
871
+ 00068/0583 7 (256,448,3)
872
+ 00068/0599 7 (256,448,3)
873
+ 00068/0739 7 (256,448,3)
874
+ 00068/0743 7 (256,448,3)
875
+ 00068/0754 7 (256,448,3)
876
+ 00068/0812 7 (256,448,3)
877
+ 00069/0178 7 (256,448,3)
878
+ 00070/0025 7 (256,448,3)
879
+ 00070/0030 7 (256,448,3)
880
+ 00070/0036 7 (256,448,3)
881
+ 00070/0042 7 (256,448,3)
882
+ 00070/0078 7 (256,448,3)
883
+ 00070/0079 7 (256,448,3)
884
+ 00070/0362 7 (256,448,3)
885
+ 00071/0195 7 (256,448,3)
886
+ 00071/0210 7 (256,448,3)
887
+ 00071/0211 7 (256,448,3)
888
+ 00071/0221 7 (256,448,3)
889
+ 00071/0352 7 (256,448,3)
890
+ 00071/0354 7 (256,448,3)
891
+ 00071/0366 7 (256,448,3)
892
+ 00071/0454 7 (256,448,3)
893
+ 00071/0464 7 (256,448,3)
894
+ 00071/0487 7 (256,448,3)
895
+ 00071/0502 7 (256,448,3)
896
+ 00071/0561 7 (256,448,3)
897
+ 00071/0676 7 (256,448,3)
898
+ 00071/0808 7 (256,448,3)
899
+ 00071/0813 7 (256,448,3)
900
+ 00071/0836 7 (256,448,3)
901
+ 00072/0286 7 (256,448,3)
902
+ 00072/0290 7 (256,448,3)
903
+ 00072/0298 7 (256,448,3)
904
+ 00072/0302 7 (256,448,3)
905
+ 00072/0333 7 (256,448,3)
906
+ 00072/0590 7 (256,448,3)
907
+ 00072/0793 7 (256,448,3)
908
+ 00072/0803 7 (256,448,3)
909
+ 00072/0833 7 (256,448,3)
910
+ 00073/0049 7 (256,448,3)
911
+ 00073/0050 7 (256,448,3)
912
+ 00073/0388 7 (256,448,3)
913
+ 00073/0480 7 (256,448,3)
914
+ 00073/0485 7 (256,448,3)
915
+ 00073/0611 7 (256,448,3)
916
+ 00073/0616 7 (256,448,3)
917
+ 00073/0714 7 (256,448,3)
918
+ 00073/0724 7 (256,448,3)
919
+ 00073/0730 7 (256,448,3)
920
+ 00074/0034 7 (256,448,3)
921
+ 00074/0228 7 (256,448,3)
922
+ 00074/0239 7 (256,448,3)
923
+ 00074/0275 7 (256,448,3)
924
+ 00074/0527 7 (256,448,3)
925
+ 00074/0620 7 (256,448,3)
926
+ 00074/0764 7 (256,448,3)
927
+ 00074/0849 7 (256,448,3)
928
+ 00074/0893 7 (256,448,3)
929
+ 00075/0333 7 (256,448,3)
930
+ 00075/0339 7 (256,448,3)
931
+ 00075/0347 7 (256,448,3)
932
+ 00075/0399 7 (256,448,3)
933
+ 00075/0478 7 (256,448,3)
934
+ 00075/0494 7 (256,448,3)
935
+ 00075/0678 7 (256,448,3)
936
+ 00075/0688 7 (256,448,3)
937
+ 00075/0706 7 (256,448,3)
938
+ 00075/0709 7 (256,448,3)
939
+ 00075/0748 7 (256,448,3)
940
+ 00075/0769 7 (256,448,3)
941
+ 00075/0777 7 (256,448,3)
942
+ 00075/0781 7 (256,448,3)
943
+ 00076/0151 7 (256,448,3)
944
+ 00076/0159 7 (256,448,3)
945
+ 00076/0164 7 (256,448,3)
946
+ 00076/0265 7 (256,448,3)
947
+ 00076/0269 7 (256,448,3)
948
+ 00076/0433 7 (256,448,3)
949
+ 00076/0813 7 (256,448,3)
950
+ 00076/0817 7 (256,448,3)
951
+ 00076/0818 7 (256,448,3)
952
+ 00076/0827 7 (256,448,3)
953
+ 00076/0874 7 (256,448,3)
954
+ 00076/0880 7 (256,448,3)
955
+ 00076/0891 7 (256,448,3)
956
+ 00076/0894 7 (256,448,3)
957
+ 00076/0909 7 (256,448,3)
958
+ 00076/0913 7 (256,448,3)
959
+ 00076/0926 7 (256,448,3)
960
+ 00076/0962 7 (256,448,3)
961
+ 00076/0973 7 (256,448,3)
962
+ 00076/0986 7 (256,448,3)
963
+ 00077/0617 7 (256,448,3)
964
+ 00077/0623 7 (256,448,3)
965
+ 00077/0628 7 (256,448,3)
966
+ 00077/0629 7 (256,448,3)
967
+ 00077/0631 7 (256,448,3)
968
+ 00077/0639 7 (256,448,3)
969
+ 00077/0982 7 (256,448,3)
970
+ 00077/0984 7 (256,448,3)
971
+ 00077/0995 7 (256,448,3)
972
+ 00077/0998 7 (256,448,3)
973
+ 00078/0001 7 (256,448,3)
974
+ 00078/0015 7 (256,448,3)
975
+ 00078/0157 7 (256,448,3)
976
+ 00078/0161 7 (256,448,3)
977
+ 00078/0175 7 (256,448,3)
978
+ 00078/0178 7 (256,448,3)
979
+ 00078/0189 7 (256,448,3)
980
+ 00078/0192 7 (256,448,3)
981
+ 00078/0229 7 (256,448,3)
982
+ 00078/0237 7 (256,448,3)
983
+ 00078/0241 7 (256,448,3)
984
+ 00078/0249 7 (256,448,3)
985
+ 00078/0251 7 (256,448,3)
986
+ 00078/0254 7 (256,448,3)
987
+ 00078/0258 7 (256,448,3)
988
+ 00078/0311 7 (256,448,3)
989
+ 00078/0603 7 (256,448,3)
990
+ 00078/0607 7 (256,448,3)
991
+ 00078/0824 7 (256,448,3)
992
+ 00079/0045 7 (256,448,3)
993
+ 00079/0048 7 (256,448,3)
994
+ 00079/0054 7 (256,448,3)
995
+ 00080/0050 7 (256,448,3)
996
+ 00080/0488 7 (256,448,3)
997
+ 00080/0494 7 (256,448,3)
998
+ 00080/0496 7 (256,448,3)
999
+ 00080/0499 7 (256,448,3)
1000
+ 00080/0502 7 (256,448,3)
1001
+ 00080/0510 7 (256,448,3)
1002
+ 00080/0534 7 (256,448,3)
1003
+ 00080/0558 7 (256,448,3)
1004
+ 00080/0571 7 (256,448,3)
1005
+ 00080/0709 7 (256,448,3)
1006
+ 00080/0882 7 (256,448,3)
1007
+ 00081/0322 7 (256,448,3)
1008
+ 00081/0428 7 (256,448,3)
1009
+ 00081/0700 7 (256,448,3)
1010
+ 00081/0706 7 (256,448,3)
1011
+ 00081/0707 7 (256,448,3)
1012
+ 00081/0937 7 (256,448,3)
1013
+ 00082/0021 7 (256,448,3)
1014
+ 00082/0424 7 (256,448,3)
1015
+ 00082/0794 7 (256,448,3)
1016
+ 00082/0807 7 (256,448,3)
1017
+ 00082/0810 7 (256,448,3)
1018
+ 00082/0824 7 (256,448,3)
1019
+ 00083/0129 7 (256,448,3)
1020
+ 00083/0131 7 (256,448,3)
1021
+ 00083/0249 7 (256,448,3)
1022
+ 00083/0250 7 (256,448,3)
1023
+ 00083/0656 7 (256,448,3)
1024
+ 00083/0812 7 (256,448,3)
1025
+ 00083/0819 7 (256,448,3)
1026
+ 00083/0824 7 (256,448,3)
1027
+ 00083/0827 7 (256,448,3)
1028
+ 00083/0841 7 (256,448,3)
1029
+ 00083/0963 7 (256,448,3)
1030
+ 00084/0047 7 (256,448,3)
1031
+ 00084/0319 7 (256,448,3)
1032
+ 00084/0334 7 (256,448,3)
1033
+ 00084/0363 7 (256,448,3)
1034
+ 00084/0493 7 (256,448,3)
1035
+ 00084/0655 7 (256,448,3)
1036
+ 00084/0752 7 (256,448,3)
1037
+ 00084/0813 7 (256,448,3)
1038
+ 00084/0886 7 (256,448,3)
1039
+ 00084/0948 7 (256,448,3)
1040
+ 00084/0976 7 (256,448,3)
1041
+ 00085/0512 7 (256,448,3)
1042
+ 00085/0641 7 (256,448,3)
1043
+ 00085/0653 7 (256,448,3)
1044
+ 00085/0655 7 (256,448,3)
1045
+ 00085/0697 7 (256,448,3)
1046
+ 00085/0698 7 (256,448,3)
1047
+ 00085/0700 7 (256,448,3)
1048
+ 00085/0703 7 (256,448,3)
1049
+ 00085/0705 7 (256,448,3)
1050
+ 00085/0709 7 (256,448,3)
1051
+ 00085/0713 7 (256,448,3)
1052
+ 00085/0739 7 (256,448,3)
1053
+ 00085/0750 7 (256,448,3)
1054
+ 00085/0763 7 (256,448,3)
1055
+ 00085/0765 7 (256,448,3)
1056
+ 00085/0769 7 (256,448,3)
1057
+ 00085/0863 7 (256,448,3)
1058
+ 00085/0868 7 (256,448,3)
1059
+ 00085/0927 7 (256,448,3)
1060
+ 00085/0936 7 (256,448,3)
1061
+ 00085/0965 7 (256,448,3)
1062
+ 00085/0969 7 (256,448,3)
1063
+ 00085/0974 7 (256,448,3)
1064
+ 00085/0981 7 (256,448,3)
1065
+ 00085/0982 7 (256,448,3)
1066
+ 00085/1000 7 (256,448,3)
1067
+ 00086/0003 7 (256,448,3)
1068
+ 00086/0009 7 (256,448,3)
1069
+ 00086/0011 7 (256,448,3)
1070
+ 00086/0028 7 (256,448,3)
1071
+ 00086/0032 7 (256,448,3)
1072
+ 00086/0034 7 (256,448,3)
1073
+ 00086/0035 7 (256,448,3)
1074
+ 00086/0042 7 (256,448,3)
1075
+ 00086/0064 7 (256,448,3)
1076
+ 00086/0066 7 (256,448,3)
1077
+ 00086/0095 7 (256,448,3)
1078
+ 00086/0099 7 (256,448,3)
1079
+ 00086/0101 7 (256,448,3)
1080
+ 00086/0104 7 (256,448,3)
1081
+ 00086/0115 7 (256,448,3)
1082
+ 00086/0116 7 (256,448,3)
1083
+ 00086/0284 7 (256,448,3)
1084
+ 00086/0291 7 (256,448,3)
1085
+ 00086/0295 7 (256,448,3)
1086
+ 00086/0302 7 (256,448,3)
1087
+ 00086/0318 7 (256,448,3)
1088
+ 00086/0666 7 (256,448,3)
1089
+ 00086/0797 7 (256,448,3)
1090
+ 00086/0851 7 (256,448,3)
1091
+ 00086/0855 7 (256,448,3)
1092
+ 00086/0874 7 (256,448,3)
1093
+ 00086/0878 7 (256,448,3)
1094
+ 00086/0881 7 (256,448,3)
1095
+ 00086/0883 7 (256,448,3)
1096
+ 00086/0896 7 (256,448,3)
1097
+ 00086/0899 7 (256,448,3)
1098
+ 00086/0903 7 (256,448,3)
1099
+ 00086/0989 7 (256,448,3)
1100
+ 00087/0008 7 (256,448,3)
1101
+ 00087/0429 7 (256,448,3)
1102
+ 00087/0511 7 (256,448,3)
1103
+ 00088/0241 7 (256,448,3)
1104
+ 00088/0319 7 (256,448,3)
1105
+ 00088/0323 7 (256,448,3)
1106
+ 00088/0411 7 (256,448,3)
1107
+ 00088/0427 7 (256,448,3)
1108
+ 00088/0452 7 (256,448,3)
1109
+ 00088/0463 7 (256,448,3)
1110
+ 00088/0476 7 (256,448,3)
1111
+ 00088/0496 7 (256,448,3)
1112
+ 00088/0559 7 (256,448,3)
1113
+ 00089/0058 7 (256,448,3)
1114
+ 00089/0061 7 (256,448,3)
1115
+ 00089/0069 7 (256,448,3)
1116
+ 00089/0077 7 (256,448,3)
1117
+ 00089/0096 7 (256,448,3)
1118
+ 00089/0099 7 (256,448,3)
1119
+ 00089/0100 7 (256,448,3)
1120
+ 00089/0211 7 (256,448,3)
1121
+ 00089/0380 7 (256,448,3)
1122
+ 00089/0381 7 (256,448,3)
1123
+ 00089/0384 7 (256,448,3)
1124
+ 00089/0390 7 (256,448,3)
1125
+ 00089/0393 7 (256,448,3)
1126
+ 00089/0394 7 (256,448,3)
1127
+ 00089/0395 7 (256,448,3)
1128
+ 00089/0406 7 (256,448,3)
1129
+ 00089/0410 7 (256,448,3)
1130
+ 00089/0412 7 (256,448,3)
1131
+ 00089/0703 7 (256,448,3)
1132
+ 00089/0729 7 (256,448,3)
1133
+ 00089/0930 7 (256,448,3)
1134
+ 00089/0952 7 (256,448,3)
1135
+ 00090/0062 7 (256,448,3)
1136
+ 00090/0101 7 (256,448,3)
1137
+ 00090/0213 7 (256,448,3)
1138
+ 00090/0216 7 (256,448,3)
1139
+ 00090/0268 7 (256,448,3)
1140
+ 00090/0406 7 (256,448,3)
1141
+ 00090/0411 7 (256,448,3)
1142
+ 00090/0442 7 (256,448,3)
1143
+ 00090/0535 7 (256,448,3)
1144
+ 00090/0542 7 (256,448,3)
1145
+ 00090/0571 7 (256,448,3)
1146
+ 00090/0934 7 (256,448,3)
1147
+ 00090/0938 7 (256,448,3)
1148
+ 00090/0947 7 (256,448,3)
1149
+ 00091/0066 7 (256,448,3)
1150
+ 00091/0448 7 (256,448,3)
1151
+ 00091/0451 7 (256,448,3)
1152
+ 00091/0454 7 (256,448,3)
1153
+ 00091/0457 7 (256,448,3)
1154
+ 00091/0467 7 (256,448,3)
1155
+ 00091/0470 7 (256,448,3)
1156
+ 00091/0477 7 (256,448,3)
1157
+ 00091/0583 7 (256,448,3)
1158
+ 00091/0981 7 (256,448,3)
1159
+ 00091/0994 7 (256,448,3)
1160
+ 00092/0112 7 (256,448,3)
1161
+ 00092/0119 7 (256,448,3)
1162
+ 00092/0129 7 (256,448,3)
1163
+ 00092/0146 7 (256,448,3)
1164
+ 00092/0149 7 (256,448,3)
1165
+ 00092/0608 7 (256,448,3)
1166
+ 00092/0643 7 (256,448,3)
1167
+ 00092/0646 7 (256,448,3)
1168
+ 00092/0766 7 (256,448,3)
1169
+ 00092/0768 7 (256,448,3)
1170
+ 00092/0779 7 (256,448,3)
1171
+ 00093/0081 7 (256,448,3)
1172
+ 00093/0085 7 (256,448,3)
1173
+ 00093/0135 7 (256,448,3)
1174
+ 00093/0241 7 (256,448,3)
1175
+ 00093/0277 7 (256,448,3)
1176
+ 00093/0283 7 (256,448,3)
1177
+ 00093/0320 7 (256,448,3)
1178
+ 00093/0598 7 (256,448,3)
1179
+ 00094/0159 7 (256,448,3)
1180
+ 00094/0253 7 (256,448,3)
1181
+ 00094/0265 7 (256,448,3)
1182
+ 00094/0267 7 (256,448,3)
1183
+ 00094/0269 7 (256,448,3)
1184
+ 00094/0281 7 (256,448,3)
1185
+ 00094/0293 7 (256,448,3)
1186
+ 00094/0404 7 (256,448,3)
1187
+ 00094/0593 7 (256,448,3)
1188
+ 00094/0612 7 (256,448,3)
1189
+ 00094/0638 7 (256,448,3)
1190
+ 00094/0656 7 (256,448,3)
1191
+ 00094/0668 7 (256,448,3)
1192
+ 00094/0786 7 (256,448,3)
1193
+ 00094/0870 7 (256,448,3)
1194
+ 00094/0897 7 (256,448,3)
1195
+ 00094/0900 7 (256,448,3)
1196
+ 00094/0944 7 (256,448,3)
1197
+ 00094/0946 7 (256,448,3)
1198
+ 00094/0952 7 (256,448,3)
1199
+ 00094/0969 7 (256,448,3)
1200
+ 00094/0973 7 (256,448,3)
1201
+ 00094/0981 7 (256,448,3)
1202
+ 00095/0088 7 (256,448,3)
1203
+ 00095/0125 7 (256,448,3)
1204
+ 00095/0130 7 (256,448,3)
1205
+ 00095/0142 7 (256,448,3)
1206
+ 00095/0151 7 (256,448,3)
1207
+ 00095/0180 7 (256,448,3)
1208
+ 00095/0192 7 (256,448,3)
1209
+ 00095/0194 7 (256,448,3)
1210
+ 00095/0195 7 (256,448,3)
1211
+ 00095/0204 7 (256,448,3)
1212
+ 00095/0245 7 (256,448,3)
1213
+ 00095/0315 7 (256,448,3)
1214
+ 00095/0321 7 (256,448,3)
1215
+ 00095/0324 7 (256,448,3)
1216
+ 00095/0327 7 (256,448,3)
1217
+ 00095/0730 7 (256,448,3)
1218
+ 00095/0731 7 (256,448,3)
1219
+ 00095/0741 7 (256,448,3)
1220
+ 00095/0948 7 (256,448,3)
1221
+ 00096/0407 7 (256,448,3)
1222
+ 00096/0420 7 (256,448,3)
1223
+ 00096/0435 7 (256,448,3)
1224
+ 00096/0682 7 (256,448,3)
1225
+ 00096/0865 7 (256,448,3)
basicsr/data/meta_info/meta_info_Vimeo90K_test_medium_GT.txt ADDED
The diff for this file is too large to render. See raw diff
 
basicsr/data/meta_info/meta_info_Vimeo90K_test_slow_GT.txt ADDED
@@ -0,0 +1,1613 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 00001/0266 7 (256,448,3)
2
+ 00001/0268 7 (256,448,3)
3
+ 00001/0275 7 (256,448,3)
4
+ 00001/0278 7 (256,448,3)
5
+ 00001/0287 7 (256,448,3)
6
+ 00001/0291 7 (256,448,3)
7
+ 00001/0627 7 (256,448,3)
8
+ 00001/0636 7 (256,448,3)
9
+ 00001/0804 7 (256,448,3)
10
+ 00001/0837 7 (256,448,3)
11
+ 00001/0849 7 (256,448,3)
12
+ 00001/0851 7 (256,448,3)
13
+ 00001/0852 7 (256,448,3)
14
+ 00001/0986 7 (256,448,3)
15
+ 00001/0991 7 (256,448,3)
16
+ 00002/0007 7 (256,448,3)
17
+ 00002/0008 7 (256,448,3)
18
+ 00002/0016 7 (256,448,3)
19
+ 00002/0036 7 (256,448,3)
20
+ 00002/0091 7 (256,448,3)
21
+ 00002/0093 7 (256,448,3)
22
+ 00002/0209 7 (256,448,3)
23
+ 00002/0235 7 (256,448,3)
24
+ 00002/0236 7 (256,448,3)
25
+ 00002/0241 7 (256,448,3)
26
+ 00002/0466 7 (256,448,3)
27
+ 00002/0504 7 (256,448,3)
28
+ 00002/0960 7 (256,448,3)
29
+ 00002/0961 7 (256,448,3)
30
+ 00002/0964 7 (256,448,3)
31
+ 00003/0007 7 (256,448,3)
32
+ 00003/0069 7 (256,448,3)
33
+ 00003/0345 7 (256,448,3)
34
+ 00003/0347 7 (256,448,3)
35
+ 00003/0372 7 (256,448,3)
36
+ 00003/0525 7 (256,448,3)
37
+ 00003/0652 7 (256,448,3)
38
+ 00003/0667 7 (256,448,3)
39
+ 00003/0669 7 (256,448,3)
40
+ 00003/0706 7 (256,448,3)
41
+ 00003/0713 7 (256,448,3)
42
+ 00003/0721 7 (256,448,3)
43
+ 00003/0747 7 (256,448,3)
44
+ 00003/0829 7 (256,448,3)
45
+ 00003/0916 7 (256,448,3)
46
+ 00003/0918 7 (256,448,3)
47
+ 00003/0924 7 (256,448,3)
48
+ 00003/0926 7 (256,448,3)
49
+ 00003/0927 7 (256,448,3)
50
+ 00004/0288 7 (256,448,3)
51
+ 00004/0303 7 (256,448,3)
52
+ 00004/0307 7 (256,448,3)
53
+ 00004/0628 7 (256,448,3)
54
+ 00004/0713 7 (256,448,3)
55
+ 00004/0715 7 (256,448,3)
56
+ 00004/0719 7 (256,448,3)
57
+ 00004/0727 7 (256,448,3)
58
+ 00004/0821 7 (256,448,3)
59
+ 00005/0006 7 (256,448,3)
60
+ 00005/0007 7 (256,448,3)
61
+ 00005/0012 7 (256,448,3)
62
+ 00005/0013 7 (256,448,3)
63
+ 00005/0040 7 (256,448,3)
64
+ 00005/0055 7 (256,448,3)
65
+ 00005/0119 7 (256,448,3)
66
+ 00005/0130 7 (256,448,3)
67
+ 00005/0185 7 (256,448,3)
68
+ 00005/0198 7 (256,448,3)
69
+ 00005/0270 7 (256,448,3)
70
+ 00005/0541 7 (256,448,3)
71
+ 00005/0560 7 (256,448,3)
72
+ 00005/0660 7 (256,448,3)
73
+ 00005/0682 7 (256,448,3)
74
+ 00005/0683 7 (256,448,3)
75
+ 00005/0688 7 (256,448,3)
76
+ 00005/0706 7 (256,448,3)
77
+ 00005/0728 7 (256,448,3)
78
+ 00005/0732 7 (256,448,3)
79
+ 00005/0739 7 (256,448,3)
80
+ 00005/0804 7 (256,448,3)
81
+ 00005/0805 7 (256,448,3)
82
+ 00005/0827 7 (256,448,3)
83
+ 00005/0828 7 (256,448,3)
84
+ 00005/0857 7 (256,448,3)
85
+ 00005/0861 7 (256,448,3)
86
+ 00005/0862 7 (256,448,3)
87
+ 00005/0868 7 (256,448,3)
88
+ 00005/0872 7 (256,448,3)
89
+ 00005/0933 7 (256,448,3)
90
+ 00005/0958 7 (256,448,3)
91
+ 00005/0960 7 (256,448,3)
92
+ 00006/0087 7 (256,448,3)
93
+ 00006/0090 7 (256,448,3)
94
+ 00006/0351 7 (256,448,3)
95
+ 00006/0353 7 (256,448,3)
96
+ 00006/0558 7 (256,448,3)
97
+ 00006/0588 7 (256,448,3)
98
+ 00006/0619 7 (256,448,3)
99
+ 00006/0621 7 (256,448,3)
100
+ 00006/0748 7 (256,448,3)
101
+ 00006/0796 7 (256,448,3)
102
+ 00006/0805 7 (256,448,3)
103
+ 00006/0807 7 (256,448,3)
104
+ 00007/0236 7 (256,448,3)
105
+ 00007/0240 7 (256,448,3)
106
+ 00007/0243 7 (256,448,3)
107
+ 00007/0246 7 (256,448,3)
108
+ 00007/0247 7 (256,448,3)
109
+ 00007/0252 7 (256,448,3)
110
+ 00007/0322 7 (256,448,3)
111
+ 00007/0458 7 (256,448,3)
112
+ 00007/0492 7 (256,448,3)
113
+ 00007/0658 7 (256,448,3)
114
+ 00007/0717 7 (256,448,3)
115
+ 00007/0722 7 (256,448,3)
116
+ 00007/0725 7 (256,448,3)
117
+ 00007/0740 7 (256,448,3)
118
+ 00007/0748 7 (256,448,3)
119
+ 00007/0749 7 (256,448,3)
120
+ 00007/0759 7 (256,448,3)
121
+ 00007/0772 7 (256,448,3)
122
+ 00007/0783 7 (256,448,3)
123
+ 00007/0787 7 (256,448,3)
124
+ 00007/0883 7 (256,448,3)
125
+ 00008/0033 7 (256,448,3)
126
+ 00008/0035 7 (256,448,3)
127
+ 00008/0091 7 (256,448,3)
128
+ 00008/0154 7 (256,448,3)
129
+ 00008/0966 7 (256,448,3)
130
+ 00008/0987 7 (256,448,3)
131
+ 00009/0108 7 (256,448,3)
132
+ 00009/0607 7 (256,448,3)
133
+ 00009/0668 7 (256,448,3)
134
+ 00009/0683 7 (256,448,3)
135
+ 00009/0941 7 (256,448,3)
136
+ 00009/0949 7 (256,448,3)
137
+ 00009/0962 7 (256,448,3)
138
+ 00009/0972 7 (256,448,3)
139
+ 00009/0974 7 (256,448,3)
140
+ 00010/0014 7 (256,448,3)
141
+ 00010/0018 7 (256,448,3)
142
+ 00010/0043 7 (256,448,3)
143
+ 00010/0099 7 (256,448,3)
144
+ 00010/0252 7 (256,448,3)
145
+ 00010/0296 7 (256,448,3)
146
+ 00010/0413 7 (256,448,3)
147
+ 00010/0422 7 (256,448,3)
148
+ 00010/0516 7 (256,448,3)
149
+ 00010/0525 7 (256,448,3)
150
+ 00010/0556 7 (256,448,3)
151
+ 00010/0701 7 (256,448,3)
152
+ 00010/0740 7 (256,448,3)
153
+ 00010/0772 7 (256,448,3)
154
+ 00010/0831 7 (256,448,3)
155
+ 00010/0925 7 (256,448,3)
156
+ 00011/0013 7 (256,448,3)
157
+ 00011/0016 7 (256,448,3)
158
+ 00011/0017 7 (256,448,3)
159
+ 00011/0249 7 (256,448,3)
160
+ 00011/0826 7 (256,448,3)
161
+ 00011/0827 7 (256,448,3)
162
+ 00011/0831 7 (256,448,3)
163
+ 00011/0833 7 (256,448,3)
164
+ 00011/0835 7 (256,448,3)
165
+ 00011/0998 7 (256,448,3)
166
+ 00012/0023 7 (256,448,3)
167
+ 00012/0024 7 (256,448,3)
168
+ 00012/0027 7 (256,448,3)
169
+ 00012/0037 7 (256,448,3)
170
+ 00012/0444 7 (256,448,3)
171
+ 00012/0445 7 (256,448,3)
172
+ 00012/0451 7 (256,448,3)
173
+ 00012/0461 7 (256,448,3)
174
+ 00012/0521 7 (256,448,3)
175
+ 00012/0758 7 (256,448,3)
176
+ 00012/0760 7 (256,448,3)
177
+ 00012/0771 7 (256,448,3)
178
+ 00012/0903 7 (256,448,3)
179
+ 00012/0909 7 (256,448,3)
180
+ 00013/0581 7 (256,448,3)
181
+ 00013/0786 7 (256,448,3)
182
+ 00013/0789 7 (256,448,3)
183
+ 00013/0791 7 (256,448,3)
184
+ 00013/0798 7 (256,448,3)
185
+ 00013/0802 7 (256,448,3)
186
+ 00013/0820 7 (256,448,3)
187
+ 00013/0850 7 (256,448,3)
188
+ 00013/0854 7 (256,448,3)
189
+ 00013/0894 7 (256,448,3)
190
+ 00013/0919 7 (256,448,3)
191
+ 00013/0999 7 (256,448,3)
192
+ 00014/0001 7 (256,448,3)
193
+ 00014/0014 7 (256,448,3)
194
+ 00014/0018 7 (256,448,3)
195
+ 00014/0244 7 (256,448,3)
196
+ 00014/0475 7 (256,448,3)
197
+ 00014/0483 7 (256,448,3)
198
+ 00014/0680 7 (256,448,3)
199
+ 00014/0700 7 (256,448,3)
200
+ 00014/0701 7 (256,448,3)
201
+ 00014/0706 7 (256,448,3)
202
+ 00014/0712 7 (256,448,3)
203
+ 00014/0713 7 (256,448,3)
204
+ 00014/0717 7 (256,448,3)
205
+ 00014/0719 7 (256,448,3)
206
+ 00014/0728 7 (256,448,3)
207
+ 00014/0734 7 (256,448,3)
208
+ 00014/0736 7 (256,448,3)
209
+ 00014/0738 7 (256,448,3)
210
+ 00014/0742 7 (256,448,3)
211
+ 00014/0745 7 (256,448,3)
212
+ 00014/0746 7 (256,448,3)
213
+ 00014/0750 7 (256,448,3)
214
+ 00014/0769 7 (256,448,3)
215
+ 00014/0774 7 (256,448,3)
216
+ 00014/0781 7 (256,448,3)
217
+ 00014/0782 7 (256,448,3)
218
+ 00014/0852 7 (256,448,3)
219
+ 00014/0853 7 (256,448,3)
220
+ 00014/0855 7 (256,448,3)
221
+ 00014/0867 7 (256,448,3)
222
+ 00014/0876 7 (256,448,3)
223
+ 00014/0881 7 (256,448,3)
224
+ 00014/0890 7 (256,448,3)
225
+ 00014/0914 7 (256,448,3)
226
+ 00015/0033 7 (256,448,3)
227
+ 00015/0113 7 (256,448,3)
228
+ 00015/0125 7 (256,448,3)
229
+ 00015/0185 7 (256,448,3)
230
+ 00015/0194 7 (256,448,3)
231
+ 00015/0202 7 (256,448,3)
232
+ 00015/0312 7 (256,448,3)
233
+ 00015/0688 7 (256,448,3)
234
+ 00015/0698 7 (256,448,3)
235
+ 00015/0788 7 (256,448,3)
236
+ 00015/0854 7 (256,448,3)
237
+ 00015/0863 7 (256,448,3)
238
+ 00015/0864 7 (256,448,3)
239
+ 00015/0918 7 (256,448,3)
240
+ 00015/0931 7 (256,448,3)
241
+ 00016/0276 7 (256,448,3)
242
+ 00016/0301 7 (256,448,3)
243
+ 00016/0306 7 (256,448,3)
244
+ 00016/0324 7 (256,448,3)
245
+ 00016/0362 7 (256,448,3)
246
+ 00016/0364 7 (256,448,3)
247
+ 00016/0370 7 (256,448,3)
248
+ 00016/0378 7 (256,448,3)
249
+ 00016/0379 7 (256,448,3)
250
+ 00016/0402 7 (256,448,3)
251
+ 00016/0405 7 (256,448,3)
252
+ 00016/0418 7 (256,448,3)
253
+ 00016/0419 7 (256,448,3)
254
+ 00016/0435 7 (256,448,3)
255
+ 00016/0501 7 (256,448,3)
256
+ 00016/0561 7 (256,448,3)
257
+ 00016/0562 7 (256,448,3)
258
+ 00016/0569 7 (256,448,3)
259
+ 00016/0591 7 (256,448,3)
260
+ 00016/0599 7 (256,448,3)
261
+ 00016/0711 7 (256,448,3)
262
+ 00016/0713 7 (256,448,3)
263
+ 00016/0813 7 (256,448,3)
264
+ 00016/0953 7 (256,448,3)
265
+ 00016/0960 7 (256,448,3)
266
+ 00016/0961 7 (256,448,3)
267
+ 00017/0519 7 (256,448,3)
268
+ 00017/0523 7 (256,448,3)
269
+ 00017/0588 7 (256,448,3)
270
+ 00017/0608 7 (256,448,3)
271
+ 00017/0609 7 (256,448,3)
272
+ 00017/0719 7 (256,448,3)
273
+ 00017/0721 7 (256,448,3)
274
+ 00017/0727 7 (256,448,3)
275
+ 00017/0728 7 (256,448,3)
276
+ 00017/0769 7 (256,448,3)
277
+ 00017/0775 7 (256,448,3)
278
+ 00017/0787 7 (256,448,3)
279
+ 00017/0797 7 (256,448,3)
280
+ 00018/0043 7 (256,448,3)
281
+ 00018/0206 7 (256,448,3)
282
+ 00018/0209 7 (256,448,3)
283
+ 00018/0211 7 (256,448,3)
284
+ 00018/0216 7 (256,448,3)
285
+ 00018/0220 7 (256,448,3)
286
+ 00018/0221 7 (256,448,3)
287
+ 00018/0252 7 (256,448,3)
288
+ 00018/0260 7 (256,448,3)
289
+ 00018/0331 7 (256,448,3)
290
+ 00018/0333 7 (256,448,3)
291
+ 00018/0447 7 (256,448,3)
292
+ 00018/0523 7 (256,448,3)
293
+ 00019/0014 7 (256,448,3)
294
+ 00019/0015 7 (256,448,3)
295
+ 00019/0019 7 (256,448,3)
296
+ 00019/0049 7 (256,448,3)
297
+ 00019/0109 7 (256,448,3)
298
+ 00019/0114 7 (256,448,3)
299
+ 00019/0125 7 (256,448,3)
300
+ 00019/0137 7 (256,448,3)
301
+ 00019/0140 7 (256,448,3)
302
+ 00019/0148 7 (256,448,3)
303
+ 00019/0153 7 (256,448,3)
304
+ 00019/0155 7 (256,448,3)
305
+ 00019/0158 7 (256,448,3)
306
+ 00019/0159 7 (256,448,3)
307
+ 00019/0160 7 (256,448,3)
308
+ 00019/0162 7 (256,448,3)
309
+ 00019/0279 7 (256,448,3)
310
+ 00019/0282 7 (256,448,3)
311
+ 00019/0409 7 (256,448,3)
312
+ 00019/0427 7 (256,448,3)
313
+ 00019/0430 7 (256,448,3)
314
+ 00019/0545 7 (256,448,3)
315
+ 00019/0555 7 (256,448,3)
316
+ 00019/0558 7 (256,448,3)
317
+ 00019/0650 7 (256,448,3)
318
+ 00019/0681 7 (256,448,3)
319
+ 00019/0747 7 (256,448,3)
320
+ 00019/0748 7 (256,448,3)
321
+ 00019/0749 7 (256,448,3)
322
+ 00019/0752 7 (256,448,3)
323
+ 00019/0768 7 (256,448,3)
324
+ 00019/0772 7 (256,448,3)
325
+ 00019/0773 7 (256,448,3)
326
+ 00019/0777 7 (256,448,3)
327
+ 00019/0795 7 (256,448,3)
328
+ 00019/0806 7 (256,448,3)
329
+ 00019/0815 7 (256,448,3)
330
+ 00019/0840 7 (256,448,3)
331
+ 00019/0844 7 (256,448,3)
332
+ 00019/0848 7 (256,448,3)
333
+ 00019/0853 7 (256,448,3)
334
+ 00019/0863 7 (256,448,3)
335
+ 00019/0888 7 (256,448,3)
336
+ 00019/0894 7 (256,448,3)
337
+ 00019/0901 7 (256,448,3)
338
+ 00019/0995 7 (256,448,3)
339
+ 00021/0030 7 (256,448,3)
340
+ 00021/0035 7 (256,448,3)
341
+ 00021/0039 7 (256,448,3)
342
+ 00021/0041 7 (256,448,3)
343
+ 00021/0044 7 (256,448,3)
344
+ 00021/0045 7 (256,448,3)
345
+ 00021/0264 7 (256,448,3)
346
+ 00021/0330 7 (256,448,3)
347
+ 00021/0332 7 (256,448,3)
348
+ 00021/0333 7 (256,448,3)
349
+ 00021/0336 7 (256,448,3)
350
+ 00021/0337 7 (256,448,3)
351
+ 00021/0338 7 (256,448,3)
352
+ 00021/0343 7 (256,448,3)
353
+ 00021/0472 7 (256,448,3)
354
+ 00021/0667 7 (256,448,3)
355
+ 00021/0731 7 (256,448,3)
356
+ 00021/0779 7 (256,448,3)
357
+ 00021/0805 7 (256,448,3)
358
+ 00021/0814 7 (256,448,3)
359
+ 00021/0818 7 (256,448,3)
360
+ 00021/0874 7 (256,448,3)
361
+ 00022/0008 7 (256,448,3)
362
+ 00022/0010 7 (256,448,3)
363
+ 00022/0231 7 (256,448,3)
364
+ 00022/0323 7 (256,448,3)
365
+ 00022/0337 7 (256,448,3)
366
+ 00022/0359 7 (256,448,3)
367
+ 00022/0377 7 (256,448,3)
368
+ 00022/0378 7 (256,448,3)
369
+ 00022/0385 7 (256,448,3)
370
+ 00022/0393 7 (256,448,3)
371
+ 00022/0424 7 (256,448,3)
372
+ 00022/0582 7 (256,448,3)
373
+ 00022/0583 7 (256,448,3)
374
+ 00022/0605 7 (256,448,3)
375
+ 00022/0632 7 (256,448,3)
376
+ 00022/0633 7 (256,448,3)
377
+ 00022/0666 7 (256,448,3)
378
+ 00022/0671 7 (256,448,3)
379
+ 00022/0673 7 (256,448,3)
380
+ 00022/0702 7 (256,448,3)
381
+ 00022/0852 7 (256,448,3)
382
+ 00022/0853 7 (256,448,3)
383
+ 00022/0971 7 (256,448,3)
384
+ 00023/0037 7 (256,448,3)
385
+ 00023/0224 7 (256,448,3)
386
+ 00023/0308 7 (256,448,3)
387
+ 00023/0393 7 (256,448,3)
388
+ 00023/0633 7 (256,448,3)
389
+ 00023/0637 7 (256,448,3)
390
+ 00023/0638 7 (256,448,3)
391
+ 00023/0770 7 (256,448,3)
392
+ 00023/0786 7 (256,448,3)
393
+ 00023/0898 7 (256,448,3)
394
+ 00024/0247 7 (256,448,3)
395
+ 00024/0251 7 (256,448,3)
396
+ 00024/0267 7 (256,448,3)
397
+ 00024/0288 7 (256,448,3)
398
+ 00024/0530 7 (256,448,3)
399
+ 00024/0569 7 (256,448,3)
400
+ 00024/0587 7 (256,448,3)
401
+ 00024/0730 7 (256,448,3)
402
+ 00024/0736 7 (256,448,3)
403
+ 00024/0742 7 (256,448,3)
404
+ 00024/0832 7 (256,448,3)
405
+ 00024/0936 7 (256,448,3)
406
+ 00025/0044 7 (256,448,3)
407
+ 00025/0047 7 (256,448,3)
408
+ 00025/0540 7 (256,448,3)
409
+ 00025/0552 7 (256,448,3)
410
+ 00025/0554 7 (256,448,3)
411
+ 00025/0559 7 (256,448,3)
412
+ 00025/0572 7 (256,448,3)
413
+ 00025/0576 7 (256,448,3)
414
+ 00025/0699 7 (256,448,3)
415
+ 00025/0709 7 (256,448,3)
416
+ 00025/0743 7 (256,448,3)
417
+ 00025/0767 7 (256,448,3)
418
+ 00025/0780 7 (256,448,3)
419
+ 00025/0782 7 (256,448,3)
420
+ 00025/0784 7 (256,448,3)
421
+ 00025/0791 7 (256,448,3)
422
+ 00025/0889 7 (256,448,3)
423
+ 00025/0890 7 (256,448,3)
424
+ 00025/0894 7 (256,448,3)
425
+ 00025/0896 7 (256,448,3)
426
+ 00025/0898 7 (256,448,3)
427
+ 00025/0905 7 (256,448,3)
428
+ 00025/0999 7 (256,448,3)
429
+ 00026/0003 7 (256,448,3)
430
+ 00026/0005 7 (256,448,3)
431
+ 00026/0011 7 (256,448,3)
432
+ 00026/0017 7 (256,448,3)
433
+ 00026/0036 7 (256,448,3)
434
+ 00026/0129 7 (256,448,3)
435
+ 00026/0131 7 (256,448,3)
436
+ 00026/0161 7 (256,448,3)
437
+ 00026/0177 7 (256,448,3)
438
+ 00026/0178 7 (256,448,3)
439
+ 00026/0180 7 (256,448,3)
440
+ 00026/0298 7 (256,448,3)
441
+ 00026/0307 7 (256,448,3)
442
+ 00026/0308 7 (256,448,3)
443
+ 00026/0312 7 (256,448,3)
444
+ 00026/0352 7 (256,448,3)
445
+ 00026/0440 7 (256,448,3)
446
+ 00026/0706 7 (256,448,3)
447
+ 00026/0708 7 (256,448,3)
448
+ 00026/0715 7 (256,448,3)
449
+ 00026/0769 7 (256,448,3)
450
+ 00026/0777 7 (256,448,3)
451
+ 00026/0779 7 (256,448,3)
452
+ 00026/0789 7 (256,448,3)
453
+ 00026/0924 7 (256,448,3)
454
+ 00026/0928 7 (256,448,3)
455
+ 00026/0932 7 (256,448,3)
456
+ 00026/0935 7 (256,448,3)
457
+ 00027/0118 7 (256,448,3)
458
+ 00027/0121 7 (256,448,3)
459
+ 00027/0155 7 (256,448,3)
460
+ 00027/0168 7 (256,448,3)
461
+ 00027/0196 7 (256,448,3)
462
+ 00027/0289 7 (256,448,3)
463
+ 00027/0294 7 (256,448,3)
464
+ 00027/0803 7 (256,448,3)
465
+ 00028/0016 7 (256,448,3)
466
+ 00028/0045 7 (256,448,3)
467
+ 00028/0063 7 (256,448,3)
468
+ 00028/0601 7 (256,448,3)
469
+ 00028/0638 7 (256,448,3)
470
+ 00028/0733 7 (256,448,3)
471
+ 00028/0736 7 (256,448,3)
472
+ 00028/0741 7 (256,448,3)
473
+ 00028/0753 7 (256,448,3)
474
+ 00028/0770 7 (256,448,3)
475
+ 00028/0771 7 (256,448,3)
476
+ 00028/0777 7 (256,448,3)
477
+ 00028/0950 7 (256,448,3)
478
+ 00028/0951 7 (256,448,3)
479
+ 00029/0048 7 (256,448,3)
480
+ 00029/0060 7 (256,448,3)
481
+ 00029/0362 7 (256,448,3)
482
+ 00029/0399 7 (256,448,3)
483
+ 00029/0404 7 (256,448,3)
484
+ 00029/0412 7 (256,448,3)
485
+ 00029/0416 7 (256,448,3)
486
+ 00029/0418 7 (256,448,3)
487
+ 00029/0428 7 (256,448,3)
488
+ 00030/0131 7 (256,448,3)
489
+ 00030/0135 7 (256,448,3)
490
+ 00030/0150 7 (256,448,3)
491
+ 00030/0245 7 (256,448,3)
492
+ 00030/0339 7 (256,448,3)
493
+ 00030/0472 7 (256,448,3)
494
+ 00030/0482 7 (256,448,3)
495
+ 00030/0500 7 (256,448,3)
496
+ 00030/0501 7 (256,448,3)
497
+ 00030/0697 7 (256,448,3)
498
+ 00030/0707 7 (256,448,3)
499
+ 00030/0733 7 (256,448,3)
500
+ 00030/0743 7 (256,448,3)
501
+ 00030/0747 7 (256,448,3)
502
+ 00030/0754 7 (256,448,3)
503
+ 00030/0755 7 (256,448,3)
504
+ 00030/0759 7 (256,448,3)
505
+ 00030/0762 7 (256,448,3)
506
+ 00030/0764 7 (256,448,3)
507
+ 00030/0767 7 (256,448,3)
508
+ 00030/0794 7 (256,448,3)
509
+ 00030/0796 7 (256,448,3)
510
+ 00030/0799 7 (256,448,3)
511
+ 00030/0814 7 (256,448,3)
512
+ 00030/0823 7 (256,448,3)
513
+ 00030/0829 7 (256,448,3)
514
+ 00030/0833 7 (256,448,3)
515
+ 00030/0848 7 (256,448,3)
516
+ 00030/0853 7 (256,448,3)
517
+ 00030/0861 7 (256,448,3)
518
+ 00031/0182 7 (256,448,3)
519
+ 00031/0275 7 (256,448,3)
520
+ 00031/0279 7 (256,448,3)
521
+ 00031/0555 7 (256,448,3)
522
+ 00031/0648 7 (256,448,3)
523
+ 00031/0663 7 (256,448,3)
524
+ 00031/0680 7 (256,448,3)
525
+ 00031/0880 7 (256,448,3)
526
+ 00031/0922 7 (256,448,3)
527
+ 00031/0925 7 (256,448,3)
528
+ 00031/0928 7 (256,448,3)
529
+ 00032/0025 7 (256,448,3)
530
+ 00032/0377 7 (256,448,3)
531
+ 00032/0378 7 (256,448,3)
532
+ 00032/0382 7 (256,448,3)
533
+ 00032/0384 7 (256,448,3)
534
+ 00032/0386 7 (256,448,3)
535
+ 00032/0389 7 (256,448,3)
536
+ 00032/0391 7 (256,448,3)
537
+ 00032/0393 7 (256,448,3)
538
+ 00032/0492 7 (256,448,3)
539
+ 00032/0497 7 (256,448,3)
540
+ 00032/0505 7 (256,448,3)
541
+ 00032/0523 7 (256,448,3)
542
+ 00032/0542 7 (256,448,3)
543
+ 00032/0544 7 (256,448,3)
544
+ 00032/0712 7 (256,448,3)
545
+ 00032/0847 7 (256,448,3)
546
+ 00032/0850 7 (256,448,3)
547
+ 00032/0875 7 (256,448,3)
548
+ 00033/0062 7 (256,448,3)
549
+ 00033/0063 7 (256,448,3)
550
+ 00033/0098 7 (256,448,3)
551
+ 00033/0101 7 (256,448,3)
552
+ 00033/0105 7 (256,448,3)
553
+ 00033/0114 7 (256,448,3)
554
+ 00033/0432 7 (256,448,3)
555
+ 00033/0441 7 (256,448,3)
556
+ 00033/0606 7 (256,448,3)
557
+ 00033/0611 7 (256,448,3)
558
+ 00033/0634 7 (256,448,3)
559
+ 00033/0787 7 (256,448,3)
560
+ 00033/0792 7 (256,448,3)
561
+ 00033/0802 7 (256,448,3)
562
+ 00033/0825 7 (256,448,3)
563
+ 00033/0835 7 (256,448,3)
564
+ 00034/0249 7 (256,448,3)
565
+ 00034/0253 7 (256,448,3)
566
+ 00034/0254 7 (256,448,3)
567
+ 00034/0282 7 (256,448,3)
568
+ 00034/0318 7 (256,448,3)
569
+ 00034/0319 7 (256,448,3)
570
+ 00034/0323 7 (256,448,3)
571
+ 00034/0336 7 (256,448,3)
572
+ 00034/0348 7 (256,448,3)
573
+ 00034/0356 7 (256,448,3)
574
+ 00034/0379 7 (256,448,3)
575
+ 00034/0387 7 (256,448,3)
576
+ 00034/0575 7 (256,448,3)
577
+ 00034/0608 7 (256,448,3)
578
+ 00034/0663 7 (256,448,3)
579
+ 00034/0811 7 (256,448,3)
580
+ 00034/0812 7 (256,448,3)
581
+ 00034/0946 7 (256,448,3)
582
+ 00034/0948 7 (256,448,3)
583
+ 00034/0950 7 (256,448,3)
584
+ 00035/0204 7 (256,448,3)
585
+ 00035/0243 7 (256,448,3)
586
+ 00035/0308 7 (256,448,3)
587
+ 00035/0465 7 (256,448,3)
588
+ 00035/0478 7 (256,448,3)
589
+ 00035/0523 7 (256,448,3)
590
+ 00035/0540 7 (256,448,3)
591
+ 00035/0544 7 (256,448,3)
592
+ 00035/0556 7 (256,448,3)
593
+ 00035/0568 7 (256,448,3)
594
+ 00035/0570 7 (256,448,3)
595
+ 00035/0609 7 (256,448,3)
596
+ 00035/0643 7 (256,448,3)
597
+ 00035/0644 7 (256,448,3)
598
+ 00035/0645 7 (256,448,3)
599
+ 00035/0646 7 (256,448,3)
600
+ 00035/0650 7 (256,448,3)
601
+ 00035/0661 7 (256,448,3)
602
+ 00035/0724 7 (256,448,3)
603
+ 00035/0725 7 (256,448,3)
604
+ 00035/0850 7 (256,448,3)
605
+ 00035/0863 7 (256,448,3)
606
+ 00035/0870 7 (256,448,3)
607
+ 00035/0951 7 (256,448,3)
608
+ 00036/0038 7 (256,448,3)
609
+ 00036/0062 7 (256,448,3)
610
+ 00036/0423 7 (256,448,3)
611
+ 00036/0737 7 (256,448,3)
612
+ 00036/0750 7 (256,448,3)
613
+ 00036/0751 7 (256,448,3)
614
+ 00036/0754 7 (256,448,3)
615
+ 00036/0929 7 (256,448,3)
616
+ 00037/0085 7 (256,448,3)
617
+ 00037/0113 7 (256,448,3)
618
+ 00037/0130 7 (256,448,3)
619
+ 00037/0153 7 (256,448,3)
620
+ 00037/0169 7 (256,448,3)
621
+ 00037/0263 7 (256,448,3)
622
+ 00037/0272 7 (256,448,3)
623
+ 00037/0273 7 (256,448,3)
624
+ 00037/0275 7 (256,448,3)
625
+ 00037/0280 7 (256,448,3)
626
+ 00037/0399 7 (256,448,3)
627
+ 00037/0456 7 (256,448,3)
628
+ 00037/0853 7 (256,448,3)
629
+ 00037/0855 7 (256,448,3)
630
+ 00037/0856 7 (256,448,3)
631
+ 00037/0857 7 (256,448,3)
632
+ 00037/0925 7 (256,448,3)
633
+ 00037/0947 7 (256,448,3)
634
+ 00038/0148 7 (256,448,3)
635
+ 00038/0533 7 (256,448,3)
636
+ 00038/0534 7 (256,448,3)
637
+ 00038/0560 7 (256,448,3)
638
+ 00038/0562 7 (256,448,3)
639
+ 00038/0566 7 (256,448,3)
640
+ 00038/0578 7 (256,448,3)
641
+ 00038/0652 7 (256,448,3)
642
+ 00038/0674 7 (256,448,3)
643
+ 00038/0685 7 (256,448,3)
644
+ 00038/0686 7 (256,448,3)
645
+ 00038/0692 7 (256,448,3)
646
+ 00038/0736 7 (256,448,3)
647
+ 00039/0035 7 (256,448,3)
648
+ 00039/0105 7 (256,448,3)
649
+ 00039/0109 7 (256,448,3)
650
+ 00039/0121 7 (256,448,3)
651
+ 00039/0128 7 (256,448,3)
652
+ 00039/0129 7 (256,448,3)
653
+ 00039/0132 7 (256,448,3)
654
+ 00039/0137 7 (256,448,3)
655
+ 00039/0157 7 (256,448,3)
656
+ 00039/0496 7 (256,448,3)
657
+ 00039/0502 7 (256,448,3)
658
+ 00039/0526 7 (256,448,3)
659
+ 00039/0529 7 (256,448,3)
660
+ 00039/0682 7 (256,448,3)
661
+ 00039/0690 7 (256,448,3)
662
+ 00039/0693 7 (256,448,3)
663
+ 00039/0703 7 (256,448,3)
664
+ 00039/0725 7 (256,448,3)
665
+ 00039/0734 7 (256,448,3)
666
+ 00040/0518 7 (256,448,3)
667
+ 00040/0728 7 (256,448,3)
668
+ 00040/0774 7 (256,448,3)
669
+ 00040/0812 7 (256,448,3)
670
+ 00040/0818 7 (256,448,3)
671
+ 00040/0827 7 (256,448,3)
672
+ 00040/0914 7 (256,448,3)
673
+ 00040/0917 7 (256,448,3)
674
+ 00040/0918 7 (256,448,3)
675
+ 00040/0924 7 (256,448,3)
676
+ 00040/0925 7 (256,448,3)
677
+ 00041/0004 7 (256,448,3)
678
+ 00041/0006 7 (256,448,3)
679
+ 00041/0013 7 (256,448,3)
680
+ 00041/0059 7 (256,448,3)
681
+ 00041/0110 7 (256,448,3)
682
+ 00041/0291 7 (256,448,3)
683
+ 00041/0366 7 (256,448,3)
684
+ 00041/0388 7 (256,448,3)
685
+ 00041/0434 7 (256,448,3)
686
+ 00041/0436 7 (256,448,3)
687
+ 00041/0450 7 (256,448,3)
688
+ 00041/0457 7 (256,448,3)
689
+ 00041/0460 7 (256,448,3)
690
+ 00041/0468 7 (256,448,3)
691
+ 00041/0471 7 (256,448,3)
692
+ 00041/0474 7 (256,448,3)
693
+ 00041/0809 7 (256,448,3)
694
+ 00041/0844 7 (256,448,3)
695
+ 00041/0858 7 (256,448,3)
696
+ 00041/0874 7 (256,448,3)
697
+ 00041/0876 7 (256,448,3)
698
+ 00042/0020 7 (256,448,3)
699
+ 00042/0205 7 (256,448,3)
700
+ 00042/0206 7 (256,448,3)
701
+ 00042/0432 7 (256,448,3)
702
+ 00042/0563 7 (256,448,3)
703
+ 00042/0569 7 (256,448,3)
704
+ 00042/0575 7 (256,448,3)
705
+ 00042/0576 7 (256,448,3)
706
+ 00042/0888 7 (256,448,3)
707
+ 00042/0892 7 (256,448,3)
708
+ 00042/0943 7 (256,448,3)
709
+ 00042/0944 7 (256,448,3)
710
+ 00043/0126 7 (256,448,3)
711
+ 00043/0130 7 (256,448,3)
712
+ 00043/0136 7 (256,448,3)
713
+ 00043/0233 7 (256,448,3)
714
+ 00043/0235 7 (256,448,3)
715
+ 00043/0237 7 (256,448,3)
716
+ 00043/0277 7 (256,448,3)
717
+ 00043/0301 7 (256,448,3)
718
+ 00043/0302 7 (256,448,3)
719
+ 00043/0303 7 (256,448,3)
720
+ 00043/0308 7 (256,448,3)
721
+ 00043/0309 7 (256,448,3)
722
+ 00043/0314 7 (256,448,3)
723
+ 00043/0713 7 (256,448,3)
724
+ 00043/0715 7 (256,448,3)
725
+ 00043/0923 7 (256,448,3)
726
+ 00044/0095 7 (256,448,3)
727
+ 00044/0255 7 (256,448,3)
728
+ 00044/0864 7 (256,448,3)
729
+ 00044/0892 7 (256,448,3)
730
+ 00044/0898 7 (256,448,3)
731
+ 00044/0993 7 (256,448,3)
732
+ 00044/0995 7 (256,448,3)
733
+ 00044/0997 7 (256,448,3)
734
+ 00045/0001 7 (256,448,3)
735
+ 00045/0006 7 (256,448,3)
736
+ 00045/0269 7 (256,448,3)
737
+ 00045/0276 7 (256,448,3)
738
+ 00045/0280 7 (256,448,3)
739
+ 00045/0281 7 (256,448,3)
740
+ 00045/0282 7 (256,448,3)
741
+ 00045/0284 7 (256,448,3)
742
+ 00045/0550 7 (256,448,3)
743
+ 00045/0571 7 (256,448,3)
744
+ 00045/0629 7 (256,448,3)
745
+ 00045/0631 7 (256,448,3)
746
+ 00045/0659 7 (256,448,3)
747
+ 00045/0693 7 (256,448,3)
748
+ 00045/0807 7 (256,448,3)
749
+ 00045/0810 7 (256,448,3)
750
+ 00045/0826 7 (256,448,3)
751
+ 00045/0849 7 (256,448,3)
752
+ 00045/0946 7 (256,448,3)
753
+ 00045/0987 7 (256,448,3)
754
+ 00045/0990 7 (256,448,3)
755
+ 00046/0104 7 (256,448,3)
756
+ 00046/0477 7 (256,448,3)
757
+ 00046/0490 7 (256,448,3)
758
+ 00046/0491 7 (256,448,3)
759
+ 00046/0509 7 (256,448,3)
760
+ 00046/0513 7 (256,448,3)
761
+ 00046/0603 7 (256,448,3)
762
+ 00046/0723 7 (256,448,3)
763
+ 00046/0744 7 (256,448,3)
764
+ 00046/0746 7 (256,448,3)
765
+ 00046/0750 7 (256,448,3)
766
+ 00046/0852 7 (256,448,3)
767
+ 00046/0927 7 (256,448,3)
768
+ 00046/0928 7 (256,448,3)
769
+ 00046/0929 7 (256,448,3)
770
+ 00046/0931 7 (256,448,3)
771
+ 00046/0936 7 (256,448,3)
772
+ 00046/0939 7 (256,448,3)
773
+ 00046/0947 7 (256,448,3)
774
+ 00046/0948 7 (256,448,3)
775
+ 00046/0950 7 (256,448,3)
776
+ 00046/0955 7 (256,448,3)
777
+ 00046/0961 7 (256,448,3)
778
+ 00047/0023 7 (256,448,3)
779
+ 00047/0029 7 (256,448,3)
780
+ 00047/0035 7 (256,448,3)
781
+ 00047/0058 7 (256,448,3)
782
+ 00047/0061 7 (256,448,3)
783
+ 00047/0065 7 (256,448,3)
784
+ 00047/0068 7 (256,448,3)
785
+ 00047/0072 7 (256,448,3)
786
+ 00047/0074 7 (256,448,3)
787
+ 00047/0148 7 (256,448,3)
788
+ 00047/0594 7 (256,448,3)
789
+ 00047/0782 7 (256,448,3)
790
+ 00047/0787 7 (256,448,3)
791
+ 00047/0860 7 (256,448,3)
792
+ 00047/0889 7 (256,448,3)
793
+ 00047/0893 7 (256,448,3)
794
+ 00047/0894 7 (256,448,3)
795
+ 00047/0902 7 (256,448,3)
796
+ 00047/0975 7 (256,448,3)
797
+ 00047/0995 7 (256,448,3)
798
+ 00048/0033 7 (256,448,3)
799
+ 00048/0113 7 (256,448,3)
800
+ 00048/0115 7 (256,448,3)
801
+ 00048/0120 7 (256,448,3)
802
+ 00048/0129 7 (256,448,3)
803
+ 00048/0136 7 (256,448,3)
804
+ 00048/0327 7 (256,448,3)
805
+ 00048/0329 7 (256,448,3)
806
+ 00048/0341 7 (256,448,3)
807
+ 00048/0343 7 (256,448,3)
808
+ 00048/0345 7 (256,448,3)
809
+ 00048/0346 7 (256,448,3)
810
+ 00048/0355 7 (256,448,3)
811
+ 00048/0359 7 (256,448,3)
812
+ 00048/0363 7 (256,448,3)
813
+ 00048/0378 7 (256,448,3)
814
+ 00048/0386 7 (256,448,3)
815
+ 00048/0387 7 (256,448,3)
816
+ 00048/0388 7 (256,448,3)
817
+ 00048/0428 7 (256,448,3)
818
+ 00048/0439 7 (256,448,3)
819
+ 00048/0507 7 (256,448,3)
820
+ 00048/0510 7 (256,448,3)
821
+ 00048/0512 7 (256,448,3)
822
+ 00048/0514 7 (256,448,3)
823
+ 00048/0539 7 (256,448,3)
824
+ 00048/0542 7 (256,448,3)
825
+ 00048/0544 7 (256,448,3)
826
+ 00048/0631 7 (256,448,3)
827
+ 00048/0632 7 (256,448,3)
828
+ 00048/0636 7 (256,448,3)
829
+ 00048/0640 7 (256,448,3)
830
+ 00048/0644 7 (256,448,3)
831
+ 00048/0653 7 (256,448,3)
832
+ 00048/0655 7 (256,448,3)
833
+ 00048/0658 7 (256,448,3)
834
+ 00048/0667 7 (256,448,3)
835
+ 00048/0688 7 (256,448,3)
836
+ 00048/0708 7 (256,448,3)
837
+ 00049/0005 7 (256,448,3)
838
+ 00049/0074 7 (256,448,3)
839
+ 00049/0077 7 (256,448,3)
840
+ 00049/0084 7 (256,448,3)
841
+ 00049/0516 7 (256,448,3)
842
+ 00049/0800 7 (256,448,3)
843
+ 00049/0900 7 (256,448,3)
844
+ 00050/0607 7 (256,448,3)
845
+ 00050/0661 7 (256,448,3)
846
+ 00050/0665 7 (256,448,3)
847
+ 00050/0685 7 (256,448,3)
848
+ 00050/0711 7 (256,448,3)
849
+ 00051/0068 7 (256,448,3)
850
+ 00051/0069 7 (256,448,3)
851
+ 00051/0076 7 (256,448,3)
852
+ 00051/0569 7 (256,448,3)
853
+ 00051/0801 7 (256,448,3)
854
+ 00051/0927 7 (256,448,3)
855
+ 00051/0945 7 (256,448,3)
856
+ 00051/0952 7 (256,448,3)
857
+ 00051/0976 7 (256,448,3)
858
+ 00051/0985 7 (256,448,3)
859
+ 00052/0012 7 (256,448,3)
860
+ 00052/0015 7 (256,448,3)
861
+ 00052/0052 7 (256,448,3)
862
+ 00052/0056 7 (256,448,3)
863
+ 00052/0060 7 (256,448,3)
864
+ 00052/0157 7 (256,448,3)
865
+ 00052/0265 7 (256,448,3)
866
+ 00052/0788 7 (256,448,3)
867
+ 00052/0790 7 (256,448,3)
868
+ 00052/0793 7 (256,448,3)
869
+ 00052/0816 7 (256,448,3)
870
+ 00052/0824 7 (256,448,3)
871
+ 00052/0918 7 (256,448,3)
872
+ 00052/0933 7 (256,448,3)
873
+ 00052/0947 7 (256,448,3)
874
+ 00053/0232 7 (256,448,3)
875
+ 00053/0277 7 (256,448,3)
876
+ 00053/0362 7 (256,448,3)
877
+ 00053/0577 7 (256,448,3)
878
+ 00053/0609 7 (256,448,3)
879
+ 00053/0612 7 (256,448,3)
880
+ 00053/0628 7 (256,448,3)
881
+ 00053/0629 7 (256,448,3)
882
+ 00053/0633 7 (256,448,3)
883
+ 00053/0659 7 (256,448,3)
884
+ 00053/0667 7 (256,448,3)
885
+ 00053/0671 7 (256,448,3)
886
+ 00053/0797 7 (256,448,3)
887
+ 00053/0804 7 (256,448,3)
888
+ 00053/0807 7 (256,448,3)
889
+ 00053/0952 7 (256,448,3)
890
+ 00053/0970 7 (256,448,3)
891
+ 00053/0981 7 (256,448,3)
892
+ 00053/0999 7 (256,448,3)
893
+ 00054/0003 7 (256,448,3)
894
+ 00054/0013 7 (256,448,3)
895
+ 00054/0020 7 (256,448,3)
896
+ 00054/0022 7 (256,448,3)
897
+ 00054/0023 7 (256,448,3)
898
+ 00054/0044 7 (256,448,3)
899
+ 00054/0051 7 (256,448,3)
900
+ 00054/0063 7 (256,448,3)
901
+ 00054/0065 7 (256,448,3)
902
+ 00054/0145 7 (256,448,3)
903
+ 00054/0153 7 (256,448,3)
904
+ 00054/0203 7 (256,448,3)
905
+ 00054/0325 7 (256,448,3)
906
+ 00054/0445 7 (256,448,3)
907
+ 00054/0448 7 (256,448,3)
908
+ 00054/0456 7 (256,448,3)
909
+ 00054/0457 7 (256,448,3)
910
+ 00054/0519 7 (256,448,3)
911
+ 00054/0524 7 (256,448,3)
912
+ 00054/0530 7 (256,448,3)
913
+ 00054/0532 7 (256,448,3)
914
+ 00054/0535 7 (256,448,3)
915
+ 00054/0574 7 (256,448,3)
916
+ 00054/0760 7 (256,448,3)
917
+ 00054/0767 7 (256,448,3)
918
+ 00054/0837 7 (256,448,3)
919
+ 00055/0011 7 (256,448,3)
920
+ 00055/0109 7 (256,448,3)
921
+ 00055/0111 7 (256,448,3)
922
+ 00055/0117 7 (256,448,3)
923
+ 00055/0119 7 (256,448,3)
924
+ 00055/0182 7 (256,448,3)
925
+ 00055/0192 7 (256,448,3)
926
+ 00055/0193 7 (256,448,3)
927
+ 00055/0200 7 (256,448,3)
928
+ 00055/0204 7 (256,448,3)
929
+ 00055/0207 7 (256,448,3)
930
+ 00055/0212 7 (256,448,3)
931
+ 00055/0213 7 (256,448,3)
932
+ 00055/0348 7 (256,448,3)
933
+ 00055/0423 7 (256,448,3)
934
+ 00055/0427 7 (256,448,3)
935
+ 00055/0456 7 (256,448,3)
936
+ 00055/0489 7 (256,448,3)
937
+ 00055/0689 7 (256,448,3)
938
+ 00055/0753 7 (256,448,3)
939
+ 00055/0802 7 (256,448,3)
940
+ 00055/0844 7 (256,448,3)
941
+ 00055/0850 7 (256,448,3)
942
+ 00055/0982 7 (256,448,3)
943
+ 00055/0993 7 (256,448,3)
944
+ 00056/0113 7 (256,448,3)
945
+ 00056/0148 7 (256,448,3)
946
+ 00056/0151 7 (256,448,3)
947
+ 00056/0316 7 (256,448,3)
948
+ 00056/0379 7 (256,448,3)
949
+ 00056/0380 7 (256,448,3)
950
+ 00056/0385 7 (256,448,3)
951
+ 00056/0505 7 (256,448,3)
952
+ 00056/0579 7 (256,448,3)
953
+ 00057/0254 7 (256,448,3)
954
+ 00057/0264 7 (256,448,3)
955
+ 00057/0272 7 (256,448,3)
956
+ 00057/0403 7 (256,448,3)
957
+ 00057/0501 7 (256,448,3)
958
+ 00057/0503 7 (256,448,3)
959
+ 00057/0884 7 (256,448,3)
960
+ 00058/0026 7 (256,448,3)
961
+ 00058/0029 7 (256,448,3)
962
+ 00058/0104 7 (256,448,3)
963
+ 00058/0124 7 (256,448,3)
964
+ 00058/0162 7 (256,448,3)
965
+ 00058/0288 7 (256,448,3)
966
+ 00058/0289 7 (256,448,3)
967
+ 00058/0323 7 (256,448,3)
968
+ 00058/0328 7 (256,448,3)
969
+ 00058/0329 7 (256,448,3)
970
+ 00058/0337 7 (256,448,3)
971
+ 00058/0367 7 (256,448,3)
972
+ 00058/0383 7 (256,448,3)
973
+ 00058/0395 7 (256,448,3)
974
+ 00060/0178 7 (256,448,3)
975
+ 00060/0182 7 (256,448,3)
976
+ 00061/0001 7 (256,448,3)
977
+ 00061/0003 7 (256,448,3)
978
+ 00061/0006 7 (256,448,3)
979
+ 00061/0443 7 (256,448,3)
980
+ 00061/0586 7 (256,448,3)
981
+ 00061/0587 7 (256,448,3)
982
+ 00061/0774 7 (256,448,3)
983
+ 00061/0789 7 (256,448,3)
984
+ 00061/0815 7 (256,448,3)
985
+ 00061/0817 7 (256,448,3)
986
+ 00061/0826 7 (256,448,3)
987
+ 00061/0829 7 (256,448,3)
988
+ 00061/0830 7 (256,448,3)
989
+ 00061/0832 7 (256,448,3)
990
+ 00061/0833 7 (256,448,3)
991
+ 00061/0836 7 (256,448,3)
992
+ 00061/0837 7 (256,448,3)
993
+ 00061/0839 7 (256,448,3)
994
+ 00061/0843 7 (256,448,3)
995
+ 00061/0849 7 (256,448,3)
996
+ 00061/0859 7 (256,448,3)
997
+ 00061/0861 7 (256,448,3)
998
+ 00061/0868 7 (256,448,3)
999
+ 00061/0877 7 (256,448,3)
1000
+ 00061/0889 7 (256,448,3)
1001
+ 00061/0905 7 (256,448,3)
1002
+ 00062/0115 7 (256,448,3)
1003
+ 00062/0118 7 (256,448,3)
1004
+ 00062/0125 7 (256,448,3)
1005
+ 00062/0134 7 (256,448,3)
1006
+ 00062/0142 7 (256,448,3)
1007
+ 00062/0400 7 (256,448,3)
1008
+ 00062/0457 7 (256,448,3)
1009
+ 00062/0459 7 (256,448,3)
1010
+ 00062/0560 7 (256,448,3)
1011
+ 00062/0650 7 (256,448,3)
1012
+ 00062/0655 7 (256,448,3)
1013
+ 00062/0715 7 (256,448,3)
1014
+ 00062/0847 7 (256,448,3)
1015
+ 00062/0905 7 (256,448,3)
1016
+ 00062/0981 7 (256,448,3)
1017
+ 00063/0177 7 (256,448,3)
1018
+ 00063/0230 7 (256,448,3)
1019
+ 00063/0253 7 (256,448,3)
1020
+ 00063/0257 7 (256,448,3)
1021
+ 00063/0326 7 (256,448,3)
1022
+ 00063/0530 7 (256,448,3)
1023
+ 00063/0677 7 (256,448,3)
1024
+ 00063/0759 7 (256,448,3)
1025
+ 00063/0761 7 (256,448,3)
1026
+ 00063/0777 7 (256,448,3)
1027
+ 00063/0842 7 (256,448,3)
1028
+ 00063/0900 7 (256,448,3)
1029
+ 00064/0014 7 (256,448,3)
1030
+ 00064/0028 7 (256,448,3)
1031
+ 00064/0029 7 (256,448,3)
1032
+ 00064/0030 7 (256,448,3)
1033
+ 00064/0037 7 (256,448,3)
1034
+ 00064/0044 7 (256,448,3)
1035
+ 00064/0280 7 (256,448,3)
1036
+ 00064/0285 7 (256,448,3)
1037
+ 00064/0286 7 (256,448,3)
1038
+ 00064/0291 7 (256,448,3)
1039
+ 00064/0300 7 (256,448,3)
1040
+ 00064/0303 7 (256,448,3)
1041
+ 00064/0308 7 (256,448,3)
1042
+ 00064/0314 7 (256,448,3)
1043
+ 00064/0316 7 (256,448,3)
1044
+ 00064/0317 7 (256,448,3)
1045
+ 00064/0323 7 (256,448,3)
1046
+ 00064/0435 7 (256,448,3)
1047
+ 00064/0733 7 (256,448,3)
1048
+ 00064/0848 7 (256,448,3)
1049
+ 00064/0868 7 (256,448,3)
1050
+ 00064/0888 7 (256,448,3)
1051
+ 00064/0898 7 (256,448,3)
1052
+ 00065/0116 7 (256,448,3)
1053
+ 00065/0121 7 (256,448,3)
1054
+ 00065/0122 7 (256,448,3)
1055
+ 00065/0124 7 (256,448,3)
1056
+ 00065/0125 7 (256,448,3)
1057
+ 00065/0126 7 (256,448,3)
1058
+ 00065/0136 7 (256,448,3)
1059
+ 00065/0146 7 (256,448,3)
1060
+ 00065/0147 7 (256,448,3)
1061
+ 00065/0163 7 (256,448,3)
1062
+ 00065/0170 7 (256,448,3)
1063
+ 00065/0175 7 (256,448,3)
1064
+ 00065/0176 7 (256,448,3)
1065
+ 00065/0180 7 (256,448,3)
1066
+ 00065/0184 7 (256,448,3)
1067
+ 00065/0186 7 (256,448,3)
1068
+ 00065/0332 7 (256,448,3)
1069
+ 00065/0343 7 (256,448,3)
1070
+ 00065/0365 7 (256,448,3)
1071
+ 00065/0393 7 (256,448,3)
1072
+ 00065/0394 7 (256,448,3)
1073
+ 00065/0442 7 (256,448,3)
1074
+ 00065/0459 7 (256,448,3)
1075
+ 00065/0462 7 (256,448,3)
1076
+ 00065/0476 7 (256,448,3)
1077
+ 00065/0483 7 (256,448,3)
1078
+ 00065/0590 7 (256,448,3)
1079
+ 00065/0593 7 (256,448,3)
1080
+ 00065/0595 7 (256,448,3)
1081
+ 00065/0774 7 (256,448,3)
1082
+ 00065/0947 7 (256,448,3)
1083
+ 00065/0985 7 (256,448,3)
1084
+ 00065/0986 7 (256,448,3)
1085
+ 00066/0015 7 (256,448,3)
1086
+ 00066/0043 7 (256,448,3)
1087
+ 00066/0131 7 (256,448,3)
1088
+ 00066/0157 7 (256,448,3)
1089
+ 00066/0169 7 (256,448,3)
1090
+ 00066/0374 7 (256,448,3)
1091
+ 00066/0382 7 (256,448,3)
1092
+ 00066/0481 7 (256,448,3)
1093
+ 00066/0482 7 (256,448,3)
1094
+ 00066/0491 7 (256,448,3)
1095
+ 00066/0493 7 (256,448,3)
1096
+ 00066/0494 7 (256,448,3)
1097
+ 00066/0496 7 (256,448,3)
1098
+ 00066/0680 7 (256,448,3)
1099
+ 00066/0700 7 (256,448,3)
1100
+ 00066/0887 7 (256,448,3)
1101
+ 00066/0910 7 (256,448,3)
1102
+ 00066/0918 7 (256,448,3)
1103
+ 00067/0024 7 (256,448,3)
1104
+ 00067/0059 7 (256,448,3)
1105
+ 00067/0408 7 (256,448,3)
1106
+ 00067/0414 7 (256,448,3)
1107
+ 00067/0417 7 (256,448,3)
1108
+ 00067/0419 7 (256,448,3)
1109
+ 00067/0423 7 (256,448,3)
1110
+ 00067/0441 7 (256,448,3)
1111
+ 00067/0467 7 (256,448,3)
1112
+ 00067/0471 7 (256,448,3)
1113
+ 00067/0487 7 (256,448,3)
1114
+ 00067/0494 7 (256,448,3)
1115
+ 00067/0497 7 (256,448,3)
1116
+ 00067/0513 7 (256,448,3)
1117
+ 00067/0521 7 (256,448,3)
1118
+ 00068/0111 7 (256,448,3)
1119
+ 00068/0123 7 (256,448,3)
1120
+ 00068/0126 7 (256,448,3)
1121
+ 00068/0129 7 (256,448,3)
1122
+ 00068/0270 7 (256,448,3)
1123
+ 00068/0330 7 (256,448,3)
1124
+ 00068/0407 7 (256,448,3)
1125
+ 00068/0428 7 (256,448,3)
1126
+ 00068/0544 7 (256,448,3)
1127
+ 00068/0635 7 (256,448,3)
1128
+ 00068/0637 7 (256,448,3)
1129
+ 00068/0736 7 (256,448,3)
1130
+ 00068/0738 7 (256,448,3)
1131
+ 00068/0747 7 (256,448,3)
1132
+ 00068/0748 7 (256,448,3)
1133
+ 00068/0749 7 (256,448,3)
1134
+ 00068/0762 7 (256,448,3)
1135
+ 00068/0815 7 (256,448,3)
1136
+ 00068/0981 7 (256,448,3)
1137
+ 00068/0982 7 (256,448,3)
1138
+ 00069/0187 7 (256,448,3)
1139
+ 00069/0191 7 (256,448,3)
1140
+ 00070/0001 7 (256,448,3)
1141
+ 00070/0003 7 (256,448,3)
1142
+ 00070/0340 7 (256,448,3)
1143
+ 00070/0341 7 (256,448,3)
1144
+ 00070/0342 7 (256,448,3)
1145
+ 00070/0347 7 (256,448,3)
1146
+ 00070/0372 7 (256,448,3)
1147
+ 00070/0383 7 (256,448,3)
1148
+ 00070/0389 7 (256,448,3)
1149
+ 00070/0728 7 (256,448,3)
1150
+ 00070/0813 7 (256,448,3)
1151
+ 00070/0814 7 (256,448,3)
1152
+ 00070/0823 7 (256,448,3)
1153
+ 00070/0840 7 (256,448,3)
1154
+ 00070/0843 7 (256,448,3)
1155
+ 00070/0861 7 (256,448,3)
1156
+ 00071/0111 7 (256,448,3)
1157
+ 00071/0138 7 (256,448,3)
1158
+ 00071/0143 7 (256,448,3)
1159
+ 00071/0150 7 (256,448,3)
1160
+ 00071/0508 7 (256,448,3)
1161
+ 00071/0514 7 (256,448,3)
1162
+ 00071/0550 7 (256,448,3)
1163
+ 00071/0556 7 (256,448,3)
1164
+ 00071/0600 7 (256,448,3)
1165
+ 00071/0665 7 (256,448,3)
1166
+ 00071/0670 7 (256,448,3)
1167
+ 00071/0672 7 (256,448,3)
1168
+ 00071/0673 7 (256,448,3)
1169
+ 00071/0705 7 (256,448,3)
1170
+ 00071/0706 7 (256,448,3)
1171
+ 00071/0707 7 (256,448,3)
1172
+ 00071/0774 7 (256,448,3)
1173
+ 00071/0799 7 (256,448,3)
1174
+ 00071/0814 7 (256,448,3)
1175
+ 00071/0816 7 (256,448,3)
1176
+ 00071/0819 7 (256,448,3)
1177
+ 00071/0823 7 (256,448,3)
1178
+ 00071/0828 7 (256,448,3)
1179
+ 00071/0830 7 (256,448,3)
1180
+ 00071/0839 7 (256,448,3)
1181
+ 00071/0841 7 (256,448,3)
1182
+ 00072/0192 7 (256,448,3)
1183
+ 00072/0194 7 (256,448,3)
1184
+ 00072/0197 7 (256,448,3)
1185
+ 00072/0199 7 (256,448,3)
1186
+ 00072/0285 7 (256,448,3)
1187
+ 00072/0586 7 (256,448,3)
1188
+ 00072/0795 7 (256,448,3)
1189
+ 00072/0811 7 (256,448,3)
1190
+ 00072/0812 7 (256,448,3)
1191
+ 00072/0824 7 (256,448,3)
1192
+ 00072/0831 7 (256,448,3)
1193
+ 00072/0835 7 (256,448,3)
1194
+ 00072/0837 7 (256,448,3)
1195
+ 00072/0841 7 (256,448,3)
1196
+ 00072/0962 7 (256,448,3)
1197
+ 00073/0296 7 (256,448,3)
1198
+ 00073/0299 7 (256,448,3)
1199
+ 00073/0300 7 (256,448,3)
1200
+ 00073/0301 7 (256,448,3)
1201
+ 00073/0427 7 (256,448,3)
1202
+ 00073/0428 7 (256,448,3)
1203
+ 00073/0494 7 (256,448,3)
1204
+ 00073/0615 7 (256,448,3)
1205
+ 00073/0620 7 (256,448,3)
1206
+ 00073/0624 7 (256,448,3)
1207
+ 00073/0979 7 (256,448,3)
1208
+ 00074/0226 7 (256,448,3)
1209
+ 00074/0250 7 (256,448,3)
1210
+ 00074/0284 7 (256,448,3)
1211
+ 00074/0503 7 (256,448,3)
1212
+ 00074/0614 7 (256,448,3)
1213
+ 00074/0629 7 (256,448,3)
1214
+ 00074/0762 7 (256,448,3)
1215
+ 00074/0765 7 (256,448,3)
1216
+ 00074/0900 7 (256,448,3)
1217
+ 00074/0908 7 (256,448,3)
1218
+ 00075/0352 7 (256,448,3)
1219
+ 00075/0360 7 (256,448,3)
1220
+ 00075/0361 7 (256,448,3)
1221
+ 00075/0365 7 (256,448,3)
1222
+ 00075/0383 7 (256,448,3)
1223
+ 00075/0384 7 (256,448,3)
1224
+ 00075/0386 7 (256,448,3)
1225
+ 00075/0407 7 (256,448,3)
1226
+ 00075/0410 7 (256,448,3)
1227
+ 00075/0412 7 (256,448,3)
1228
+ 00075/0413 7 (256,448,3)
1229
+ 00075/0459 7 (256,448,3)
1230
+ 00075/0504 7 (256,448,3)
1231
+ 00075/0515 7 (256,448,3)
1232
+ 00075/0518 7 (256,448,3)
1233
+ 00075/0567 7 (256,448,3)
1234
+ 00075/0681 7 (256,448,3)
1235
+ 00075/0693 7 (256,448,3)
1236
+ 00075/0728 7 (256,448,3)
1237
+ 00075/0731 7 (256,448,3)
1238
+ 00075/0804 7 (256,448,3)
1239
+ 00075/0974 7 (256,448,3)
1240
+ 00075/0975 7 (256,448,3)
1241
+ 00075/0983 7 (256,448,3)
1242
+ 00075/0997 7 (256,448,3)
1243
+ 00076/0006 7 (256,448,3)
1244
+ 00076/0007 7 (256,448,3)
1245
+ 00076/0011 7 (256,448,3)
1246
+ 00076/0013 7 (256,448,3)
1247
+ 00076/0014 7 (256,448,3)
1248
+ 00076/0027 7 (256,448,3)
1249
+ 00076/0029 7 (256,448,3)
1250
+ 00076/0037 7 (256,448,3)
1251
+ 00076/0041 7 (256,448,3)
1252
+ 00076/0055 7 (256,448,3)
1253
+ 00076/0071 7 (256,448,3)
1254
+ 00076/0172 7 (256,448,3)
1255
+ 00076/0275 7 (256,448,3)
1256
+ 00076/0286 7 (256,448,3)
1257
+ 00076/0467 7 (256,448,3)
1258
+ 00076/0481 7 (256,448,3)
1259
+ 00076/0527 7 (256,448,3)
1260
+ 00076/0895 7 (256,448,3)
1261
+ 00076/0896 7 (256,448,3)
1262
+ 00076/0906 7 (256,448,3)
1263
+ 00076/0924 7 (256,448,3)
1264
+ 00076/0964 7 (256,448,3)
1265
+ 00076/0984 7 (256,448,3)
1266
+ 00077/0317 7 (256,448,3)
1267
+ 00077/0322 7 (256,448,3)
1268
+ 00077/0333 7 (256,448,3)
1269
+ 00077/0334 7 (256,448,3)
1270
+ 00077/0480 7 (256,448,3)
1271
+ 00077/0488 7 (256,448,3)
1272
+ 00077/0490 7 (256,448,3)
1273
+ 00077/0582 7 (256,448,3)
1274
+ 00077/0586 7 (256,448,3)
1275
+ 00077/0969 7 (256,448,3)
1276
+ 00078/0007 7 (256,448,3)
1277
+ 00078/0011 7 (256,448,3)
1278
+ 00078/0153 7 (256,448,3)
1279
+ 00078/0289 7 (256,448,3)
1280
+ 00078/0312 7 (256,448,3)
1281
+ 00078/0492 7 (256,448,3)
1282
+ 00078/0580 7 (256,448,3)
1283
+ 00078/0595 7 (256,448,3)
1284
+ 00078/0814 7 (256,448,3)
1285
+ 00078/0950 7 (256,448,3)
1286
+ 00078/0955 7 (256,448,3)
1287
+ 00079/0060 7 (256,448,3)
1288
+ 00079/0067 7 (256,448,3)
1289
+ 00080/0216 7 (256,448,3)
1290
+ 00080/0308 7 (256,448,3)
1291
+ 00080/0504 7 (256,448,3)
1292
+ 00080/0552 7 (256,448,3)
1293
+ 00080/0576 7 (256,448,3)
1294
+ 00080/0583 7 (256,448,3)
1295
+ 00080/0837 7 (256,448,3)
1296
+ 00080/0839 7 (256,448,3)
1297
+ 00080/0871 7 (256,448,3)
1298
+ 00080/0877 7 (256,448,3)
1299
+ 00080/0880 7 (256,448,3)
1300
+ 00080/0969 7 (256,448,3)
1301
+ 00080/0973 7 (256,448,3)
1302
+ 00080/0980 7 (256,448,3)
1303
+ 00081/0202 7 (256,448,3)
1304
+ 00081/0203 7 (256,448,3)
1305
+ 00081/0210 7 (256,448,3)
1306
+ 00081/0268 7 (256,448,3)
1307
+ 00081/0281 7 (256,448,3)
1308
+ 00081/0283 7 (256,448,3)
1309
+ 00081/0317 7 (256,448,3)
1310
+ 00081/0327 7 (256,448,3)
1311
+ 00082/0018 7 (256,448,3)
1312
+ 00082/0025 7 (256,448,3)
1313
+ 00082/0089 7 (256,448,3)
1314
+ 00082/0140 7 (256,448,3)
1315
+ 00082/0442 7 (256,448,3)
1316
+ 00082/0465 7 (256,448,3)
1317
+ 00082/0473 7 (256,448,3)
1318
+ 00082/0481 7 (256,448,3)
1319
+ 00082/0492 7 (256,448,3)
1320
+ 00082/0495 7 (256,448,3)
1321
+ 00082/0497 7 (256,448,3)
1322
+ 00082/0502 7 (256,448,3)
1323
+ 00082/0504 7 (256,448,3)
1324
+ 00082/0506 7 (256,448,3)
1325
+ 00082/0507 7 (256,448,3)
1326
+ 00082/0510 7 (256,448,3)
1327
+ 00082/0519 7 (256,448,3)
1328
+ 00082/0523 7 (256,448,3)
1329
+ 00082/0588 7 (256,448,3)
1330
+ 00082/0597 7 (256,448,3)
1331
+ 00082/0632 7 (256,448,3)
1332
+ 00082/0751 7 (256,448,3)
1333
+ 00082/0767 7 (256,448,3)
1334
+ 00082/0771 7 (256,448,3)
1335
+ 00082/0790 7 (256,448,3)
1336
+ 00082/0804 7 (256,448,3)
1337
+ 00082/0823 7 (256,448,3)
1338
+ 00083/0052 7 (256,448,3)
1339
+ 00083/0056 7 (256,448,3)
1340
+ 00083/0113 7 (256,448,3)
1341
+ 00083/0114 7 (256,448,3)
1342
+ 00083/0122 7 (256,448,3)
1343
+ 00083/0137 7 (256,448,3)
1344
+ 00083/0270 7 (256,448,3)
1345
+ 00083/0295 7 (256,448,3)
1346
+ 00083/0303 7 (256,448,3)
1347
+ 00083/0308 7 (256,448,3)
1348
+ 00083/0586 7 (256,448,3)
1349
+ 00083/0592 7 (256,448,3)
1350
+ 00083/0640 7 (256,448,3)
1351
+ 00083/0648 7 (256,448,3)
1352
+ 00083/0654 7 (256,448,3)
1353
+ 00083/0662 7 (256,448,3)
1354
+ 00083/0666 7 (256,448,3)
1355
+ 00083/0668 7 (256,448,3)
1356
+ 00083/0669 7 (256,448,3)
1357
+ 00083/0675 7 (256,448,3)
1358
+ 00083/0679 7 (256,448,3)
1359
+ 00083/0681 7 (256,448,3)
1360
+ 00083/0682 7 (256,448,3)
1361
+ 00083/0694 7 (256,448,3)
1362
+ 00083/0695 7 (256,448,3)
1363
+ 00083/0697 7 (256,448,3)
1364
+ 00083/0704 7 (256,448,3)
1365
+ 00083/0713 7 (256,448,3)
1366
+ 00083/0721 7 (256,448,3)
1367
+ 00083/0855 7 (256,448,3)
1368
+ 00084/0109 7 (256,448,3)
1369
+ 00084/0113 7 (256,448,3)
1370
+ 00084/0306 7 (256,448,3)
1371
+ 00084/0442 7 (256,448,3)
1372
+ 00084/0669 7 (256,448,3)
1373
+ 00084/0679 7 (256,448,3)
1374
+ 00084/0685 7 (256,448,3)
1375
+ 00084/0691 7 (256,448,3)
1376
+ 00084/0768 7 (256,448,3)
1377
+ 00084/0817 7 (256,448,3)
1378
+ 00085/0027 7 (256,448,3)
1379
+ 00085/0035 7 (256,448,3)
1380
+ 00085/0038 7 (256,448,3)
1381
+ 00085/0223 7 (256,448,3)
1382
+ 00085/0233 7 (256,448,3)
1383
+ 00085/0281 7 (256,448,3)
1384
+ 00085/0287 7 (256,448,3)
1385
+ 00085/0313 7 (256,448,3)
1386
+ 00085/0521 7 (256,448,3)
1387
+ 00085/0848 7 (256,448,3)
1388
+ 00085/0855 7 (256,448,3)
1389
+ 00085/0865 7 (256,448,3)
1390
+ 00085/0952 7 (256,448,3)
1391
+ 00085/0964 7 (256,448,3)
1392
+ 00085/0973 7 (256,448,3)
1393
+ 00085/0986 7 (256,448,3)
1394
+ 00085/0993 7 (256,448,3)
1395
+ 00086/0070 7 (256,448,3)
1396
+ 00086/0075 7 (256,448,3)
1397
+ 00086/0094 7 (256,448,3)
1398
+ 00086/0103 7 (256,448,3)
1399
+ 00086/0112 7 (256,448,3)
1400
+ 00086/0288 7 (256,448,3)
1401
+ 00086/0576 7 (256,448,3)
1402
+ 00086/0580 7 (256,448,3)
1403
+ 00086/0584 7 (256,448,3)
1404
+ 00086/0599 7 (256,448,3)
1405
+ 00086/0600 7 (256,448,3)
1406
+ 00086/0602 7 (256,448,3)
1407
+ 00086/0612 7 (256,448,3)
1408
+ 00086/0629 7 (256,448,3)
1409
+ 00086/0655 7 (256,448,3)
1410
+ 00086/0679 7 (256,448,3)
1411
+ 00086/0694 7 (256,448,3)
1412
+ 00086/0695 7 (256,448,3)
1413
+ 00086/0701 7 (256,448,3)
1414
+ 00086/0760 7 (256,448,3)
1415
+ 00086/0786 7 (256,448,3)
1416
+ 00086/0845 7 (256,448,3)
1417
+ 00086/0868 7 (256,448,3)
1418
+ 00086/0889 7 (256,448,3)
1419
+ 00086/0891 7 (256,448,3)
1420
+ 00086/0927 7 (256,448,3)
1421
+ 00086/0938 7 (256,448,3)
1422
+ 00086/0946 7 (256,448,3)
1423
+ 00086/0963 7 (256,448,3)
1424
+ 00086/0969 7 (256,448,3)
1425
+ 00087/0023 7 (256,448,3)
1426
+ 00087/0029 7 (256,448,3)
1427
+ 00087/0144 7 (256,448,3)
1428
+ 00087/0148 7 (256,448,3)
1429
+ 00087/0159 7 (256,448,3)
1430
+ 00087/0174 7 (256,448,3)
1431
+ 00087/0283 7 (256,448,3)
1432
+ 00087/0284 7 (256,448,3)
1433
+ 00087/0294 7 (256,448,3)
1434
+ 00087/0296 7 (256,448,3)
1435
+ 00087/0498 7 (256,448,3)
1436
+ 00087/0502 7 (256,448,3)
1437
+ 00087/0532 7 (256,448,3)
1438
+ 00087/0557 7 (256,448,3)
1439
+ 00087/0559 7 (256,448,3)
1440
+ 00087/0574 7 (256,448,3)
1441
+ 00087/0577 7 (256,448,3)
1442
+ 00088/0006 7 (256,448,3)
1443
+ 00088/0268 7 (256,448,3)
1444
+ 00088/0320 7 (256,448,3)
1445
+ 00088/0412 7 (256,448,3)
1446
+ 00088/0431 7 (256,448,3)
1447
+ 00088/0432 7 (256,448,3)
1448
+ 00088/0465 7 (256,448,3)
1449
+ 00088/0507 7 (256,448,3)
1450
+ 00088/0565 7 (256,448,3)
1451
+ 00088/0629 7 (256,448,3)
1452
+ 00088/0831 7 (256,448,3)
1453
+ 00088/0836 7 (256,448,3)
1454
+ 00088/0972 7 (256,448,3)
1455
+ 00088/0974 7 (256,448,3)
1456
+ 00088/0980 7 (256,448,3)
1457
+ 00089/0067 7 (256,448,3)
1458
+ 00089/0244 7 (256,448,3)
1459
+ 00089/0404 7 (256,448,3)
1460
+ 00089/0416 7 (256,448,3)
1461
+ 00089/0419 7 (256,448,3)
1462
+ 00089/0428 7 (256,448,3)
1463
+ 00089/0712 7 (256,448,3)
1464
+ 00089/0713 7 (256,448,3)
1465
+ 00089/0723 7 (256,448,3)
1466
+ 00089/0727 7 (256,448,3)
1467
+ 00089/0770 7 (256,448,3)
1468
+ 00089/0809 7 (256,448,3)
1469
+ 00089/0811 7 (256,448,3)
1470
+ 00089/0888 7 (256,448,3)
1471
+ 00089/0898 7 (256,448,3)
1472
+ 00089/0903 7 (256,448,3)
1473
+ 00089/0907 7 (256,448,3)
1474
+ 00089/0911 7 (256,448,3)
1475
+ 00089/0915 7 (256,448,3)
1476
+ 00089/0926 7 (256,448,3)
1477
+ 00089/0955 7 (256,448,3)
1478
+ 00090/0027 7 (256,448,3)
1479
+ 00090/0028 7 (256,448,3)
1480
+ 00090/0032 7 (256,448,3)
1481
+ 00090/0038 7 (256,448,3)
1482
+ 00090/0076 7 (256,448,3)
1483
+ 00090/0081 7 (256,448,3)
1484
+ 00090/0086 7 (256,448,3)
1485
+ 00090/0119 7 (256,448,3)
1486
+ 00090/0258 7 (256,448,3)
1487
+ 00090/0261 7 (256,448,3)
1488
+ 00090/0447 7 (256,448,3)
1489
+ 00090/0498 7 (256,448,3)
1490
+ 00090/0514 7 (256,448,3)
1491
+ 00090/0523 7 (256,448,3)
1492
+ 00090/0530 7 (256,448,3)
1493
+ 00090/0540 7 (256,448,3)
1494
+ 00090/0548 7 (256,448,3)
1495
+ 00090/0565 7 (256,448,3)
1496
+ 00090/0578 7 (256,448,3)
1497
+ 00090/0580 7 (256,448,3)
1498
+ 00090/0581 7 (256,448,3)
1499
+ 00090/0780 7 (256,448,3)
1500
+ 00090/0940 7 (256,448,3)
1501
+ 00090/0984 7 (256,448,3)
1502
+ 00091/0023 7 (256,448,3)
1503
+ 00091/0051 7 (256,448,3)
1504
+ 00091/0317 7 (256,448,3)
1505
+ 00091/0320 7 (256,448,3)
1506
+ 00091/0582 7 (256,448,3)
1507
+ 00091/0585 7 (256,448,3)
1508
+ 00091/0588 7 (256,448,3)
1509
+ 00091/0601 7 (256,448,3)
1510
+ 00091/0602 7 (256,448,3)
1511
+ 00091/0603 7 (256,448,3)
1512
+ 00091/0634 7 (256,448,3)
1513
+ 00091/0693 7 (256,448,3)
1514
+ 00091/0741 7 (256,448,3)
1515
+ 00091/0966 7 (256,448,3)
1516
+ 00091/0973 7 (256,448,3)
1517
+ 00091/0985 7 (256,448,3)
1518
+ 00092/0007 7 (256,448,3)
1519
+ 00092/0132 7 (256,448,3)
1520
+ 00092/0270 7 (256,448,3)
1521
+ 00092/0296 7 (256,448,3)
1522
+ 00092/0611 7 (256,448,3)
1523
+ 00092/0625 7 (256,448,3)
1524
+ 00092/0627 7 (256,448,3)
1525
+ 00092/0651 7 (256,448,3)
1526
+ 00092/0652 7 (256,448,3)
1527
+ 00092/0910 7 (256,448,3)
1528
+ 00093/0075 7 (256,448,3)
1529
+ 00093/0078 7 (256,448,3)
1530
+ 00093/0100 7 (256,448,3)
1531
+ 00093/0132 7 (256,448,3)
1532
+ 00093/0133 7 (256,448,3)
1533
+ 00093/0176 7 (256,448,3)
1534
+ 00093/0177 7 (256,448,3)
1535
+ 00093/0178 7 (256,448,3)
1536
+ 00093/0181 7 (256,448,3)
1537
+ 00093/0183 7 (256,448,3)
1538
+ 00093/0184 7 (256,448,3)
1539
+ 00093/0286 7 (256,448,3)
1540
+ 00093/0304 7 (256,448,3)
1541
+ 00093/0305 7 (256,448,3)
1542
+ 00093/0319 7 (256,448,3)
1543
+ 00093/0324 7 (256,448,3)
1544
+ 00093/0325 7 (256,448,3)
1545
+ 00093/0327 7 (256,448,3)
1546
+ 00093/0331 7 (256,448,3)
1547
+ 00093/0444 7 (256,448,3)
1548
+ 00093/0450 7 (256,448,3)
1549
+ 00093/0593 7 (256,448,3)
1550
+ 00094/0032 7 (256,448,3)
1551
+ 00094/0057 7 (256,448,3)
1552
+ 00094/0139 7 (256,448,3)
1553
+ 00094/0206 7 (256,448,3)
1554
+ 00094/0211 7 (256,448,3)
1555
+ 00094/0215 7 (256,448,3)
1556
+ 00094/0218 7 (256,448,3)
1557
+ 00094/0257 7 (256,448,3)
1558
+ 00094/0329 7 (256,448,3)
1559
+ 00094/0331 7 (256,448,3)
1560
+ 00094/0332 7 (256,448,3)
1561
+ 00094/0369 7 (256,448,3)
1562
+ 00094/0370 7 (256,448,3)
1563
+ 00094/0383 7 (256,448,3)
1564
+ 00094/0385 7 (256,448,3)
1565
+ 00094/0387 7 (256,448,3)
1566
+ 00094/0399 7 (256,448,3)
1567
+ 00094/0605 7 (256,448,3)
1568
+ 00094/0648 7 (256,448,3)
1569
+ 00094/0649 7 (256,448,3)
1570
+ 00094/0759 7 (256,448,3)
1571
+ 00094/0800 7 (256,448,3)
1572
+ 00094/0894 7 (256,448,3)
1573
+ 00094/0896 7 (256,448,3)
1574
+ 00095/0089 7 (256,448,3)
1575
+ 00095/0108 7 (256,448,3)
1576
+ 00095/0109 7 (256,448,3)
1577
+ 00095/0114 7 (256,448,3)
1578
+ 00095/0128 7 (256,448,3)
1579
+ 00095/0133 7 (256,448,3)
1580
+ 00095/0150 7 (256,448,3)
1581
+ 00095/0153 7 (256,448,3)
1582
+ 00095/0154 7 (256,448,3)
1583
+ 00095/0196 7 (256,448,3)
1584
+ 00095/0209 7 (256,448,3)
1585
+ 00095/0228 7 (256,448,3)
1586
+ 00095/0230 7 (256,448,3)
1587
+ 00095/0231 7 (256,448,3)
1588
+ 00095/0242 7 (256,448,3)
1589
+ 00095/0243 7 (256,448,3)
1590
+ 00095/0253 7 (256,448,3)
1591
+ 00095/0280 7 (256,448,3)
1592
+ 00095/0281 7 (256,448,3)
1593
+ 00095/0283 7 (256,448,3)
1594
+ 00095/0314 7 (256,448,3)
1595
+ 00095/0868 7 (256,448,3)
1596
+ 00095/0894 7 (256,448,3)
1597
+ 00096/0062 7 (256,448,3)
1598
+ 00096/0347 7 (256,448,3)
1599
+ 00096/0348 7 (256,448,3)
1600
+ 00096/0359 7 (256,448,3)
1601
+ 00096/0363 7 (256,448,3)
1602
+ 00096/0373 7 (256,448,3)
1603
+ 00096/0378 7 (256,448,3)
1604
+ 00096/0387 7 (256,448,3)
1605
+ 00096/0395 7 (256,448,3)
1606
+ 00096/0396 7 (256,448,3)
1607
+ 00096/0404 7 (256,448,3)
1608
+ 00096/0653 7 (256,448,3)
1609
+ 00096/0668 7 (256,448,3)
1610
+ 00096/0679 7 (256,448,3)
1611
+ 00096/0729 7 (256,448,3)
1612
+ 00096/0736 7 (256,448,3)
1613
+ 00096/0823 7 (256,448,3)
basicsr/data/meta_info/meta_info_Vimeo90K_train_GT.txt ADDED
The diff for this file is too large to render. See raw diff
 
basicsr/data/paired_image_dataset.py ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from torch.utils import data as data
2
+ from torchvision.transforms.functional import normalize
3
+
4
+ from basicsr.data.data_util import paired_paths_from_folder, paired_paths_from_lmdb, paired_paths_from_meta_info_file
5
+ from basicsr.data.transforms import augment, paired_random_crop
6
+ from basicsr.utils import FileClient, bgr2ycbcr, imfrombytes, img2tensor
7
+ from basicsr.utils.registry import DATASET_REGISTRY
8
+
9
+
10
+ @DATASET_REGISTRY.register()
11
+ class PairedImageDataset(data.Dataset):
12
+ """Paired image dataset for image restoration.
13
+
14
+ Read LQ (Low Quality, e.g. LR (Low Resolution), blurry, noisy, etc) and GT image pairs.
15
+
16
+ There are three modes:
17
+
18
+ 1. **lmdb**: Use lmdb files. If opt['io_backend'] == lmdb.
19
+ 2. **meta_info_file**: Use meta information file to generate paths. \
20
+ If opt['io_backend'] != lmdb and opt['meta_info_file'] is not None.
21
+ 3. **folder**: Scan folders to generate paths. The rest.
22
+
23
+ Args:
24
+ opt (dict): Config for train datasets. It contains the following keys:
25
+ dataroot_gt (str): Data root path for gt.
26
+ dataroot_lq (str): Data root path for lq.
27
+ meta_info_file (str): Path for meta information file.
28
+ io_backend (dict): IO backend type and other kwarg.
29
+ filename_tmpl (str): Template for each filename. Note that the template excludes the file extension.
30
+ Default: '{}'.
31
+ gt_size (int): Cropped patched size for gt patches.
32
+ use_hflip (bool): Use horizontal flips.
33
+ use_rot (bool): Use rotation (use vertical flip and transposing h and w for implementation).
34
+ scale (bool): Scale, which will be added automatically.
35
+ phase (str): 'train' or 'val'.
36
+ """
37
+
38
+ def __init__(self, opt):
39
+ super(PairedImageDataset, self).__init__()
40
+ self.opt = opt
41
+ # file client (io backend)
42
+ self.file_client = None
43
+ self.io_backend_opt = opt['io_backend']
44
+ self.mean = opt['mean'] if 'mean' in opt else None
45
+ self.std = opt['std'] if 'std' in opt else None
46
+
47
+ self.gt_folder, self.lq_folder = opt['dataroot_gt'], opt['dataroot_lq']
48
+ if 'filename_tmpl' in opt:
49
+ self.filename_tmpl = opt['filename_tmpl']
50
+ else:
51
+ self.filename_tmpl = '{}'
52
+
53
+ if self.io_backend_opt['type'] == 'lmdb':
54
+ self.io_backend_opt['db_paths'] = [self.lq_folder, self.gt_folder]
55
+ self.io_backend_opt['client_keys'] = ['lq', 'gt']
56
+ self.paths = paired_paths_from_lmdb([self.lq_folder, self.gt_folder], ['lq', 'gt'])
57
+ elif 'meta_info_file' in self.opt and self.opt['meta_info_file'] is not None:
58
+ self.paths = paired_paths_from_meta_info_file([self.lq_folder, self.gt_folder], ['lq', 'gt'],
59
+ self.opt['meta_info_file'], self.filename_tmpl)
60
+ else:
61
+ self.paths = paired_paths_from_folder([self.lq_folder, self.gt_folder], ['lq', 'gt'], self.filename_tmpl)
62
+
63
+ def __getitem__(self, index):
64
+ if self.file_client is None:
65
+ self.file_client = FileClient(self.io_backend_opt.pop('type'), **self.io_backend_opt)
66
+
67
+ scale = self.opt['scale']
68
+
69
+ # Load gt and lq images. Dimension order: HWC; channel order: BGR;
70
+ # image range: [0, 1], float32.
71
+ gt_path = self.paths[index]['gt_path']
72
+ img_bytes = self.file_client.get(gt_path, 'gt')
73
+ img_gt = imfrombytes(img_bytes, float32=True)
74
+ lq_path = self.paths[index]['lq_path']
75
+ img_bytes = self.file_client.get(lq_path, 'lq')
76
+ img_lq = imfrombytes(img_bytes, float32=True)
77
+
78
+ # augmentation for training
79
+ if self.opt['phase'] == 'train':
80
+ gt_size = self.opt['gt_size']
81
+ # random crop
82
+ img_gt, img_lq = paired_random_crop(img_gt, img_lq, gt_size, scale, gt_path)
83
+ # flip, rotation
84
+ img_gt, img_lq = augment([img_gt, img_lq], self.opt['use_hflip'], self.opt['use_rot'])
85
+
86
+ # color space transform
87
+ if 'color' in self.opt and self.opt['color'] == 'y':
88
+ img_gt = bgr2ycbcr(img_gt, y_only=True)[..., None]
89
+ img_lq = bgr2ycbcr(img_lq, y_only=True)[..., None]
90
+
91
+ # crop the unmatched GT images during validation or testing, especially for SR benchmark datasets
92
+ # TODO: It is better to update the datasets, rather than force to crop
93
+ if self.opt['phase'] != 'train':
94
+ img_gt = img_gt[0:img_lq.shape[0] * scale, 0:img_lq.shape[1] * scale, :]
95
+
96
+ # BGR to RGB, HWC to CHW, numpy to tensor
97
+ img_gt, img_lq = img2tensor([img_gt, img_lq], bgr2rgb=True, float32=True)
98
+ # normalize
99
+ if self.mean is not None or self.std is not None:
100
+ normalize(img_lq, self.mean, self.std, inplace=True)
101
+ normalize(img_gt, self.mean, self.std, inplace=True)
102
+
103
+ return {'lq': img_lq, 'gt': img_gt, 'lq_path': lq_path, 'gt_path': gt_path}
104
+
105
+ def __len__(self):
106
+ return len(self.paths)
basicsr/data/prefetch_dataloader.py ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import queue as Queue
2
+ import threading
3
+ import torch
4
+ from torch.utils.data import DataLoader
5
+
6
+
7
+ class PrefetchGenerator(threading.Thread):
8
+ """A general prefetch generator.
9
+
10
+ Reference: https://stackoverflow.com/questions/7323664/python-generator-pre-fetch
11
+
12
+ Args:
13
+ generator: Python generator.
14
+ num_prefetch_queue (int): Number of prefetch queue.
15
+ """
16
+
17
+ def __init__(self, generator, num_prefetch_queue):
18
+ threading.Thread.__init__(self)
19
+ self.queue = Queue.Queue(num_prefetch_queue)
20
+ self.generator = generator
21
+ self.daemon = True
22
+ self.start()
23
+
24
+ def run(self):
25
+ for item in self.generator:
26
+ self.queue.put(item)
27
+ self.queue.put(None)
28
+
29
+ def __next__(self):
30
+ next_item = self.queue.get()
31
+ if next_item is None:
32
+ raise StopIteration
33
+ return next_item
34
+
35
+ def __iter__(self):
36
+ return self
37
+
38
+
39
+ class PrefetchDataLoader(DataLoader):
40
+ """Prefetch version of dataloader.
41
+
42
+ Reference: https://github.com/IgorSusmelj/pytorch-styleguide/issues/5#
43
+
44
+ TODO:
45
+ Need to test on single gpu and ddp (multi-gpu). There is a known issue in
46
+ ddp.
47
+
48
+ Args:
49
+ num_prefetch_queue (int): Number of prefetch queue.
50
+ kwargs (dict): Other arguments for dataloader.
51
+ """
52
+
53
+ def __init__(self, num_prefetch_queue, **kwargs):
54
+ self.num_prefetch_queue = num_prefetch_queue
55
+ super(PrefetchDataLoader, self).__init__(**kwargs)
56
+
57
+ def __iter__(self):
58
+ return PrefetchGenerator(super().__iter__(), self.num_prefetch_queue)
59
+
60
+
61
+ class CPUPrefetcher():
62
+ """CPU prefetcher.
63
+
64
+ Args:
65
+ loader: Dataloader.
66
+ """
67
+
68
+ def __init__(self, loader):
69
+ self.ori_loader = loader
70
+ self.loader = iter(loader)
71
+
72
+ def next(self):
73
+ try:
74
+ return next(self.loader)
75
+ except StopIteration:
76
+ return None
77
+
78
+ def reset(self):
79
+ self.loader = iter(self.ori_loader)
80
+
81
+
82
+ class CUDAPrefetcher():
83
+ """CUDA prefetcher.
84
+
85
+ Reference: https://github.com/NVIDIA/apex/issues/304#
86
+
87
+ It may consume more GPU memory.
88
+
89
+ Args:
90
+ loader: Dataloader.
91
+ opt (dict): Options.
92
+ """
93
+
94
+ def __init__(self, loader, opt):
95
+ self.ori_loader = loader
96
+ self.loader = iter(loader)
97
+ self.opt = opt
98
+ self.stream = torch.cuda.Stream()
99
+ self.device = torch.device('cuda' if opt['num_gpu'] != 0 else 'cpu')
100
+ self.preload()
101
+
102
+ def preload(self):
103
+ try:
104
+ self.batch = next(self.loader) # self.batch is a dict
105
+ except StopIteration:
106
+ self.batch = None
107
+ return None
108
+ # put tensors to gpu
109
+ with torch.cuda.stream(self.stream):
110
+ for k, v in self.batch.items():
111
+ if torch.is_tensor(v):
112
+ self.batch[k] = self.batch[k].to(device=self.device, non_blocking=True)
113
+
114
+ def next(self):
115
+ torch.cuda.current_stream().wait_stream(self.stream)
116
+ batch = self.batch
117
+ self.preload()
118
+ return batch
119
+
120
+ def reset(self):
121
+ self.loader = iter(self.ori_loader)
122
+ self.preload()
basicsr/data/realesrgan_dataset.py ADDED
@@ -0,0 +1,384 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import cv2
2
+ import math
3
+ import numpy as np
4
+ import os
5
+ import os.path as osp
6
+ import random
7
+ import time
8
+ import torch
9
+ from pathlib import Path
10
+
11
+ import albumentations
12
+
13
+ import torch.nn.functional as F
14
+ from torch.utils import data as data
15
+
16
+ from basicsr.utils import DiffJPEG
17
+ from basicsr.data.degradations import circular_lowpass_kernel, random_mixed_kernels
18
+ from basicsr.data.transforms import augment
19
+ from basicsr.utils import FileClient, get_root_logger, imfrombytes, img2tensor
20
+ from basicsr.utils.registry import DATASET_REGISTRY
21
+ from basicsr.utils.img_process_util import filter2D
22
+ from basicsr.data.transforms import paired_random_crop, random_crop
23
+ from basicsr.data.degradations import random_add_gaussian_noise_pt, random_add_poisson_noise_pt
24
+
25
+ from utils import util_image
26
+
27
+ def readline_txt(txt_file):
28
+ txt_file = [txt_file, ] if isinstance(txt_file, str) else txt_file
29
+ out = []
30
+ for txt_file_current in txt_file:
31
+ with open(txt_file_current, 'r') as ff:
32
+ out.extend([x[:-1] for x in ff.readlines()])
33
+
34
+ return out
35
+
36
+ @DATASET_REGISTRY.register(suffix='basicsr')
37
+ class RealESRGANDataset(data.Dataset):
38
+ """Dataset used for Real-ESRGAN model:
39
+ Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data.
40
+
41
+ It loads gt (Ground-Truth) images, and augments them.
42
+ It also generates blur kernels and sinc kernels for generating low-quality images.
43
+ Note that the low-quality images are processed in tensors on GPUS for faster processing.
44
+
45
+ Args:
46
+ opt (dict): Config for train datasets. It contains the following keys:
47
+ dataroot_gt (str): Data root path for gt.
48
+ meta_info (str): Path for meta information file.
49
+ io_backend (dict): IO backend type and other kwarg.
50
+ use_hflip (bool): Use horizontal flips.
51
+ use_rot (bool): Use rotation (use vertical flip and transposing h and w for implementation).
52
+ Please see more options in the codes.
53
+ """
54
+
55
+ def __init__(self, opt, mode='training'):
56
+ super(RealESRGANDataset, self).__init__()
57
+ self.opt = opt
58
+ self.file_client = None
59
+ self.io_backend_opt = opt['io_backend']
60
+
61
+ # file client (lmdb io backend)
62
+ self.image_paths = []
63
+ self.text_paths = []
64
+ self.moment_paths = []
65
+ if opt.get('data_source', None) is not None:
66
+ for ii in range(len(opt['data_source'])):
67
+ configs = opt['data_source'].get(f'source{ii+1}')
68
+ root_path = Path(configs.root_path)
69
+ im_folder = root_path / configs.image_path
70
+ im_ext = configs.im_ext
71
+ image_stems = sorted([x.stem for x in im_folder.glob(f"*.{im_ext}")])
72
+ if configs.get('length', None) is not None:
73
+ assert configs.length < len(image_stems)
74
+ image_stems = image_stems[:configs.length]
75
+
76
+ if configs.get("text_path", None) is not None:
77
+ text_folder = root_path / configs.text_path
78
+ text_stems = [x.stem for x in text_folder.glob("*.txt")]
79
+ image_stems = sorted(list(set(image_stems).intersection(set(text_stems))))
80
+ self.text_paths.extend([str(text_folder / f"{x}.txt") for x in image_stems])
81
+ else:
82
+ self.text_paths.extend([None, ] * len(image_stems))
83
+
84
+ self.image_paths.extend([str(im_folder / f"{x}.{im_ext}") for x in image_stems])
85
+
86
+ if configs.get("moment_path", None) is not None:
87
+ moment_folder = root_path / configs.moment_path
88
+ self.moment_paths.extend([str(moment_folder / f"{x}.npy") for x in image_stems])
89
+ else:
90
+ self.moment_paths.extend([None, ] * len(image_stems))
91
+
92
+ # blur settings for the first degradation
93
+ self.blur_kernel_size = opt['blur_kernel_size']
94
+ self.kernel_list = opt['kernel_list']
95
+ self.kernel_prob = opt['kernel_prob'] # a list for each kernel probability
96
+ self.blur_sigma = opt['blur_sigma']
97
+ self.betag_range = opt['betag_range'] # betag used in generalized Gaussian blur kernels
98
+ self.betap_range = opt['betap_range'] # betap used in plateau blur kernels
99
+ self.sinc_prob = opt['sinc_prob'] # the probability for sinc filters
100
+
101
+ # blur settings for the second degradation
102
+ self.blur_kernel_size2 = opt['blur_kernel_size2']
103
+ self.kernel_list2 = opt['kernel_list2']
104
+ self.kernel_prob2 = opt['kernel_prob2']
105
+ self.blur_sigma2 = opt['blur_sigma2']
106
+ self.betag_range2 = opt['betag_range2']
107
+ self.betap_range2 = opt['betap_range2']
108
+ self.sinc_prob2 = opt['sinc_prob2']
109
+
110
+ # a final sinc filter
111
+ self.final_sinc_prob = opt['final_sinc_prob']
112
+
113
+ self.kernel_range1 = [x for x in range(3, opt['blur_kernel_size'], 2)] # kernel size ranges from 7 to 21
114
+ self.kernel_range2 = [x for x in range(3, opt['blur_kernel_size2'], 2)] # kernel size ranges from 7 to 21
115
+ # TODO: kernel range is now hard-coded, should be in the configure file
116
+ # convolving with pulse tensor brings no blurry effect
117
+ self.pulse_tensor = torch.zeros(opt['blur_kernel_size2'], opt['blur_kernel_size2']).float()
118
+ self.pulse_tensor[opt['blur_kernel_size2']//2, opt['blur_kernel_size2']//2] = 1
119
+
120
+ self.mode = mode
121
+
122
+ def __getitem__(self, index):
123
+ if self.file_client is None:
124
+ self.file_client = FileClient(self.io_backend_opt.pop('type'), **self.io_backend_opt)
125
+
126
+ # -------------------------------- Load gt images -------------------------------- #
127
+ # Shape: (h, w, c); channel order: BGR; image range: [0, 1], float32.
128
+ gt_path = self.image_paths[index]
129
+ # avoid errors caused by high latency in reading files
130
+ retry = 3
131
+ while retry > 0:
132
+ try:
133
+ img_bytes = self.file_client.get(gt_path, 'gt')
134
+ img_gt = imfrombytes(img_bytes, float32=True)
135
+ except:
136
+ index = random.randint(0, self.__len__())
137
+ gt_path = self.image_paths[index]
138
+ time.sleep(1) # sleep 1s for occasional server congestion
139
+ finally:
140
+ retry -= 1
141
+ if self.mode == 'testing':
142
+ if not hasattr(self, 'test_aug'):
143
+ self.test_aug = albumentations.Compose([
144
+ albumentations.SmallestMaxSize(
145
+ max_size=self.opt['gt_size'],
146
+ interpolation=cv2.INTER_AREA,
147
+ ),
148
+ albumentations.CenterCrop(self.opt['gt_size'], self.opt['gt_size']),
149
+ ])
150
+ img_gt = self.test_aug(image=img_gt)['image']
151
+ elif self.mode == 'training':
152
+ # -------------------- Do augmentation for training: flip, rotation -------------------- #
153
+ if self.opt['use_hflip'] or self.opt['use_rot']:
154
+ img_gt = augment(img_gt, self.opt['use_hflip'], self.opt['use_rot'])
155
+
156
+ h, w = img_gt.shape[0:2]
157
+ gt_size = self.opt['gt_size']
158
+
159
+ # resize or pad
160
+ if not self.opt['random_crop']:
161
+ if not min(h, w) == gt_size:
162
+ if not hasattr(self, 'smallest_resizer'):
163
+ self.smallest_resizer = util_image.SmallestMaxSize(
164
+ max_size=gt_size, pass_resize=False,
165
+ )
166
+ img_gt = self.smallest_resizer(img_gt)
167
+
168
+ # center crop
169
+ if not hasattr(self, 'center_cropper'):
170
+ self.center_cropper = albumentations.CenterCrop(gt_size, gt_size)
171
+ img_gt = self.center_cropper(image=img_gt)['image']
172
+ else:
173
+ img_gt = random_crop(img_gt, self.opt['gt_size'])
174
+ else:
175
+ raise ValueError(f'Unexpected value {self.mode} for mode parameter')
176
+
177
+ # ------------------------ Generate kernels (used in the first degradation) ------------------------ #
178
+ kernel_size = random.choice(self.kernel_range1)
179
+ if np.random.uniform() < self.opt['sinc_prob']:
180
+ # this sinc filter setting is for kernels ranging from [7, 21]
181
+ if kernel_size < 13:
182
+ omega_c = np.random.uniform(np.pi / 3, np.pi)
183
+ else:
184
+ omega_c = np.random.uniform(np.pi / 5, np.pi)
185
+ kernel = circular_lowpass_kernel(omega_c, kernel_size, pad_to=False)
186
+ else:
187
+ kernel = random_mixed_kernels(
188
+ self.kernel_list,
189
+ self.kernel_prob,
190
+ kernel_size,
191
+ self.blur_sigma,
192
+ self.blur_sigma, [-math.pi, math.pi],
193
+ self.betag_range,
194
+ self.betap_range,
195
+ noise_range=None)
196
+ # pad kernel
197
+ pad_size = (self.blur_kernel_size - kernel_size) // 2
198
+ kernel = np.pad(kernel, ((pad_size, pad_size), (pad_size, pad_size)))
199
+
200
+ # ------------------------ Generate kernels (used in the second degradation) ------------------------ #
201
+ kernel_size = random.choice(self.kernel_range2)
202
+ if np.random.uniform() < self.opt['sinc_prob2']:
203
+ if kernel_size < 13:
204
+ omega_c = np.random.uniform(np.pi / 3, np.pi)
205
+ else:
206
+ omega_c = np.random.uniform(np.pi / 5, np.pi)
207
+ kernel2 = circular_lowpass_kernel(omega_c, kernel_size, pad_to=False)
208
+ else:
209
+ kernel2 = random_mixed_kernels(
210
+ self.kernel_list2,
211
+ self.kernel_prob2,
212
+ kernel_size,
213
+ self.blur_sigma2,
214
+ self.blur_sigma2, [-math.pi, math.pi],
215
+ self.betag_range2,
216
+ self.betap_range2,
217
+ noise_range=None)
218
+
219
+ # pad kernel
220
+ pad_size = (self.blur_kernel_size2 - kernel_size) // 2
221
+ kernel2 = np.pad(kernel2, ((pad_size, pad_size), (pad_size, pad_size)))
222
+
223
+ # ------------------------------------- the final sinc kernel ------------------------------------- #
224
+ if np.random.uniform() < self.opt['final_sinc_prob']:
225
+ kernel_size = random.choice(self.kernel_range2)
226
+ omega_c = np.random.uniform(np.pi / 3, np.pi)
227
+ sinc_kernel = circular_lowpass_kernel(omega_c, kernel_size, pad_to=self.blur_kernel_size2)
228
+ sinc_kernel = torch.FloatTensor(sinc_kernel)
229
+ else:
230
+ sinc_kernel = self.pulse_tensor
231
+
232
+ # BGR to RGB, HWC to CHW, numpy to tensor
233
+ img_gt = img2tensor([img_gt], bgr2rgb=True, float32=True)[0]
234
+ kernel = torch.FloatTensor(kernel)
235
+ kernel2 = torch.FloatTensor(kernel2)
236
+
237
+ if self.text_paths[index] is None or self.opt['random_crop']:
238
+ prompt = ""
239
+ else:
240
+ with open(self.text_paths[index], 'r') as ff:
241
+ prompt = ff.read()
242
+ if self.opt.max_token_length is not None:
243
+ prompt = prompt[:self.opt.max_token_length]
244
+
245
+ return_d = {
246
+ 'gt': img_gt,
247
+ 'gt_path': gt_path,
248
+ 'txt': prompt,
249
+ 'kernel1': kernel,
250
+ 'kernel2': kernel2,
251
+ 'sinc_kernel': sinc_kernel,
252
+ }
253
+ if self.moment_paths[index] is not None and (not self.opt['random_crop']):
254
+ return_d['gt_moment'] = np.load(self.moment_paths[index])
255
+
256
+ return return_d
257
+
258
+ def __len__(self):
259
+ return len(self.image_paths)
260
+
261
+ def degrade_fun(self, conf_degradation, im_gt, kernel1, kernel2, sinc_kernel):
262
+ if not hasattr(self, 'jpeger'):
263
+ self.jpeger = DiffJPEG(differentiable=False) # simulate JPEG compression artifacts
264
+
265
+ ori_h, ori_w = im_gt.size()[2:4]
266
+ sf = conf_degradation.sf
267
+
268
+ # ----------------------- The first degradation process ----------------------- #
269
+ # blur
270
+ out = filter2D(im_gt, kernel1)
271
+ # random resize
272
+ updown_type = random.choices(
273
+ ['up', 'down', 'keep'],
274
+ conf_degradation['resize_prob'],
275
+ )[0]
276
+ if updown_type == 'up':
277
+ scale = random.uniform(1, conf_degradation['resize_range'][1])
278
+ elif updown_type == 'down':
279
+ scale = random.uniform(conf_degradation['resize_range'][0], 1)
280
+ else:
281
+ scale = 1
282
+ mode = random.choice(['area', 'bilinear', 'bicubic'])
283
+ out = F.interpolate(out, scale_factor=scale, mode=mode)
284
+ # add noise
285
+ gray_noise_prob = conf_degradation['gray_noise_prob']
286
+ if random.random() < conf_degradation['gaussian_noise_prob']:
287
+ out = random_add_gaussian_noise_pt(
288
+ out,
289
+ sigma_range=conf_degradation['noise_range'],
290
+ clip=True,
291
+ rounds=False,
292
+ gray_prob=gray_noise_prob,
293
+ )
294
+ else:
295
+ out = random_add_poisson_noise_pt(
296
+ out,
297
+ scale_range=conf_degradation['poisson_scale_range'],
298
+ gray_prob=gray_noise_prob,
299
+ clip=True,
300
+ rounds=False)
301
+ # JPEG compression
302
+ jpeg_p = out.new_zeros(out.size(0)).uniform_(*conf_degradation['jpeg_range'])
303
+ out = torch.clamp(out, 0, 1) # clamp to [0, 1], otherwise JPEGer will result in unpleasant artifacts
304
+ out = self.jpeger(out, quality=jpeg_p)
305
+
306
+ # ----------------------- The second degradation process ----------------------- #
307
+ # blur
308
+ if random.random() < conf_degradation['second_order_prob']:
309
+ if random.random() < conf_degradation['second_blur_prob']:
310
+ out = filter2D(out, kernel2)
311
+ # random resize
312
+ updown_type = random.choices(
313
+ ['up', 'down', 'keep'],
314
+ conf_degradation['resize_prob2'],
315
+ )[0]
316
+ if updown_type == 'up':
317
+ scale = random.uniform(1, conf_degradation['resize_range2'][1])
318
+ elif updown_type == 'down':
319
+ scale = random.uniform(conf_degradation['resize_range2'][0], 1)
320
+ else:
321
+ scale = 1
322
+ mode = random.choice(['area', 'bilinear', 'bicubic'])
323
+ out = F.interpolate(
324
+ out,
325
+ size=(int(ori_h / sf * scale), int(ori_w / sf * scale)),
326
+ mode=mode,
327
+ )
328
+ # add noise
329
+ gray_noise_prob = conf_degradation['gray_noise_prob2']
330
+ if random.random() < conf_degradation['gaussian_noise_prob2']:
331
+ out = random_add_gaussian_noise_pt(
332
+ out,
333
+ sigma_range=conf_degradation['noise_range2'],
334
+ clip=True,
335
+ rounds=False,
336
+ gray_prob=gray_noise_prob,
337
+ )
338
+ else:
339
+ out = random_add_poisson_noise_pt(
340
+ out,
341
+ scale_range=conf_degradation['poisson_scale_range2'],
342
+ gray_prob=gray_noise_prob,
343
+ clip=True,
344
+ rounds=False,
345
+ )
346
+
347
+ # JPEG compression + the final sinc filter
348
+ # We also need to resize images to desired sizes. We group [resize back + sinc filter] together
349
+ # as one operation.
350
+ # We consider two orders:
351
+ # 1. [resize back + sinc filter] + JPEG compression
352
+ # 2. JPEG compression + [resize back + sinc filter]
353
+ # Empirically, we find other combinations (sinc + JPEG + Resize) will introduce twisted lines.
354
+ if random.random() < 0.5:
355
+ # resize back + the final sinc filter
356
+ mode = random.choice(['area', 'bilinear', 'bicubic'])
357
+ out = F.interpolate(
358
+ out,
359
+ size=(ori_h // sf, ori_w // sf),
360
+ mode=mode,
361
+ )
362
+ out = filter2D(out, sinc_kernel)
363
+ # JPEG compression
364
+ jpeg_p = out.new_zeros(out.size(0)).uniform_(*conf_degradation['jpeg_range2'])
365
+ out = torch.clamp(out, 0, 1)
366
+ out = self.jpeger(out, quality=jpeg_p)
367
+ else:
368
+ # JPEG compression
369
+ jpeg_p = out.new_zeros(out.size(0)).uniform_(*conf_degradation['jpeg_range2'])
370
+ out = torch.clamp(out, 0, 1)
371
+ out = self.jpeger(out, quality=jpeg_p)
372
+ # resize back + the final sinc filter
373
+ mode = random.choice(['area', 'bilinear', 'bicubic'])
374
+ out = F.interpolate(
375
+ out,
376
+ size=(ori_h // sf, ori_w // sf),
377
+ mode=mode,
378
+ )
379
+ out = filter2D(out, sinc_kernel)
380
+
381
+ # clamp and round
382
+ im_lq = torch.clamp((out * 255.0).round(), 0, 255) / 255.
383
+
384
+ return {'lq':im_lq.contiguous(), 'gt':im_gt}