update

b6c19984 · dengjb · b6c19984 · b6c19984 · b6c19984 · b6c19984
Commit b6c19984 authored Nov 18, 2025 by dengjb
20 changed files
--- a/projects/NAIC20/configs/r34-ibn.yml
+++ b/projects/NAIC20/configs/r34-ibn.yml
+_BASE_: Base-naic.yml
+MODEL:
+  BACKBONE:
+    NAME: build_resnet_backbone
+    DEPTH: 34x
+    FEAT_DIM: 512
+    WITH_IBN: True
+    PRETRAIN: True
+OUTPUT_DIR: projects/NAIC20/logs/r34_ibn-128x256
\ No newline at end of file
--- a/projects/NAIC20/configs/submit.yml
+++ b/projects/NAIC20/configs/submit.yml
+_BASE_: Base-naic.yml
+MODEL:
+  BACKBONE:
+    NAME: build_resnet_backbone
+    DEPTH: 34x
+    FEAT_DIM: 512
+    WITH_IBN: True
+  WEIGHTS: projects/NAIC20/logs/reproduce/r34-tripletx10/model_best.pth
+DATASETS:
+  TESTS: ("NAIC20_R2A",)
+TEST:
+  RERANK:
+    ENABLED: True
+    K1: 20
+    K2: 3
+    LAMBDA: 0.8
+  FLIP:
+    ENABLED: True
+  SAVE_DISTMAT: True
+OUTPUT_DIR: projects/NAIC20/logs/r34_ibn-128x256-submit
\ No newline at end of file
--- a/projects/NAIC20/label.txt
+++ b/projects/NAIC20/label.txt
--- a/projects/NAIC20/naic/__init__.py
+++ b/projects/NAIC20/naic/__init__.py
+# encoding: utf-8
+"""
+@author:  l1aoxingyu
+@contact: sherlockliao01@gmail.com
+"""
+from .naic_dataset import *
+from .config import add_naic_config
+from .naic_evaluator import NaicEvaluator
--- a/projects/NAIC20/naic/config.py
+++ b/projects/NAIC20/naic/config.py
+# encoding: utf-8
+"""
+@author:  xingyu liao
+@contact: sherlockliao01@gmail.com
+"""
+def add_naic_config(cfg):
+    _C = cfg
+    _C.DATASETS.RM_LT = True
+    _C.TEST.SAVE_DISTMAT = False
--- a/projects/NAIC20/naic/naic_dataset.py
+++ b/projects/NAIC20/naic/naic_dataset.py
+# encoding: utf-8
+"""
+@author:  xingyu liao
+@contact: sherlockliao01@gmail.com
+"""
+import glob
+import os
+from collections import defaultdict
+from fastreid.data.datasets import DATASET_REGISTRY
+from fastreid.data.datasets.bases import ImageDataset
+__all__ = ["NAIC20_R2", "NAIC20_R2CNV", "NAIC20_R1", "NAIC20_R1CNV", "NAIC19", "NAIC20_R2A", ]
+@DATASET_REGISTRY.register()
+class NAIC20_R2(ImageDataset):
+    dataset_name = "naic20_r2"
+    dataset_dir = "naic/2020_NAIC/fusai/train"
+    def __init__(self, root="datasets", rm_lt=False, **kwargs):
+        self.root = root
+        self.data_path = os.path.join(self.root, self.dataset_dir, "images")
+        self.train_label = os.path.join(self.root, self.dataset_dir, "naic20r2_train_list_clean.txt")
+        self.query_label = os.path.join(self.root, self.dataset_dir, "val_query.txt")
+        self.gallery_label = os.path.join(self.root, self.dataset_dir, "val_gallery.txt")
+        required_files = [self.train_label, self.query_label, self.gallery_label]
+        self.check_before_run(required_files)
+        all_train = self.process_train(self.train_label)
+        # fmt: off
+        if rm_lt: train = self.remove_longtail(all_train)
+        else:     train = all_train
+        # fmt: on
+        query, gallery = self.process_test(self.query_label, self.gallery_label)
+        super().__init__(train, query, gallery, **kwargs)
+    def process_train(self, label_path):
+        with open(label_path, 'r') as f:
+            data_list = [i.strip('\n') for i in f.readlines()]
+        img_paths = []
+        for data_info in data_list:
+            img_name, pid = data_info.split(":")
+            img_path = os.path.join(self.data_path, img_name)
+            pid = self.dataset_name + "_" + pid
+            camid = self.dataset_name + '_0'
+            img_paths.append([img_path, pid, camid])
+        return img_paths
+    def process_test(self, query_path, gallery_path):
+        with open(query_path, 'r') as f:
+            query_list = [i.strip('\n') for i in f.readlines()]
+        with open(gallery_path, 'r') as f:
+            gallery_list = [i.strip('\n') for i in f.readlines()]
+        query_paths = []
+        for data in query_list:
+            img_name, pid = data.split(':')
+            img_path = os.path.join(self.data_path, img_name)
+            camid = '0'
+            query_paths.append([img_path, int(pid), camid])
+        gallery_paths = []
+        for data in gallery_list:
+            img_name, pid = data.split(':')
+            img_path = os.path.join(self.data_path, img_name)
+            camid = '1'
+            gallery_paths.append([img_path, int(pid), camid])
+        return query_paths, gallery_paths
+    @classmethod
+    def remove_longtail(cls, all_train):
+        # 建立 id 到 image 的字典
+        pid2data = defaultdict(list)
+        for item in all_train:
+            pid2data[item[1]].append(item)
+        train = []
+        for pid, data in pid2data.items():
+            # 如果 id 只有一张图片，去掉这个 id
+            if len(data) == 1: continue
+            train.extend(data)
+        return train
+@DATASET_REGISTRY.register()
+class NAIC20_R2CNV(NAIC20_R2, ImageDataset):
+    dataset_name = 'naic20_r2cnv'
+    dataset_dir = "naic/2020_NAIC/fusai/train"
+    def __init__(self, root="datasets", rm_lt=False, **kwargs):
+        self.root = root
+        self.data_path = os.path.join(self.root, self.dataset_dir, "images_convert")
+        self.train_label = os.path.join(self.root, self.dataset_dir, "naic20r2_train_list_clean.txt")
+        self.query_label = os.path.join(self.root, self.dataset_dir, "val_query.txt")
+        self.gallery_label = os.path.join(self.root, self.dataset_dir, "val_gallery.txt")
+        required_files = [self.train_label, self.query_label, self.gallery_label]
+        self.check_before_run(required_files)
+        all_train = self.process_train(self.train_label)[:53000]
+        # fmt: off
+        if rm_lt: train = self.remove_longtail(all_train)
+        else:     train = all_train
+        # fmt: on
+        ImageDataset.__init__(self, train, query=[], gallery=[], **kwargs)
+@DATASET_REGISTRY.register()
+class NAIC20_R1(NAIC20_R2):
+    dataset_name = "naic20_r1"
+    dataset_dir = 'naic/2020_NAIC/chusai/train'
+    def __init__(self, root="datasets", rm_lt=False, **kwargs):
+        self.root = root
+        self.data_path = os.path.join(self.root, self.dataset_dir, "images")
+        self.train_label = os.path.join(self.root, self.dataset_dir, "label.txt")
+        required_files = [self.train_label]
+        self.check_before_run(required_files)
+        all_train = self.process_train(self.train_label)[:40188]
+        # fmt: off
+        if rm_lt: train = self.remove_longtail(all_train)
+        else:     train = all_train
+        # fmt: on
+        super(NAIC20_R2, self).__init__(train, [], [], **kwargs)
+@DATASET_REGISTRY.register()
+class NAIC20_R1CNV(NAIC20_R2):
+    dataset_name = 'naic20_r1cnv'
+    dataset_dir = "naic/2020_NAIC/chusai/train"
+    def __init__(self, root="datasets", rm_lt=False, **kwargs):
+        self.root = root
+        self.data_path = os.path.join(self.root, self.dataset_dir, "images_convert")
+        self.train_label = os.path.join(self.root, self.dataset_dir, "label.txt")
+        required_files = [self.train_label]
+        self.check_before_run(required_files)
+        all_train = self.process_train(self.train_label)[:40188]
+        # fmt: off
+        if rm_lt: train = self.remove_longtail(all_train)
+        else:     train = all_train
+        # fmt: on
+        super(NAIC20_R2, self).__init__(train, [], [], **kwargs)
+@DATASET_REGISTRY.register()
+class NAIC19(NAIC20_R2):
+    dataset_name = "naic19"
+    dataset_dir = "naic/2019_NAIC/fusai"
+    def __init__(self, root='datasets', rm_lt=False, **kwargs):
+        self.root = root
+        self.data_path = os.path.join(self.root, self.dataset_dir)
+        self.train_label = os.path.join(self.root, self.dataset_dir, 'train_list_clean.txt')
+        required_files = [self.train_label]
+        self.check_before_run(required_files)
+        all_train = self.process_train(self.train_label)
+        # fmt: off
+        if rm_lt: train = self.remove_longtail(all_train)
+        else:     train = all_train
+        # fmt: on
+        super(NAIC20_R2, self).__init__(train, [], [], **kwargs)
+    def process_train(self, label_path):
+        with open(label_path, 'r') as f:
+            data_list = [i.strip('\n') for i in f.readlines()]
+        img_paths = []
+        for data_info in data_list:
+            img_name, pid = data_info.split(" ")
+            img_path = os.path.join(self.data_path, img_name)
+            pid = self.dataset_name + "_" + pid
+            camid = self.dataset_name + '_0'
+            img_paths.append([img_path, pid, camid])
+        return img_paths
+@DATASET_REGISTRY.register()
+class NAIC20_R2A(ImageDataset):
+    dataset_name = "naic20_b"
+    dataset_dir = 'naic/round2/image_A'
+    def __init__(self, root='datasets', **kwargs):
+        self.root = root
+        self.query_path = os.path.join(self.root, self.dataset_dir, "query")
+        self.gallery_path = os.path.join(self.root, self.dataset_dir, "gallery")
+        query = self.process_test(self.query_path)
+        gallery = self.process_test(self.gallery_path)
+        super().__init__([], query, gallery)
+    def process_test(self, test_path):
+        img_paths = glob.glob(os.path.join(test_path, "*.png"))
+        data = []
+        for img_path in img_paths:
+            img_name = img_path.split("/")[-1]
+            data.append([img_path, img_name, "naic_0"])
+        return data
--- a/projects/NAIC20/naic/naic_evaluator.py
+++ b/projects/NAIC20/naic/naic_evaluator.py
+# encoding: utf-8
+"""
+@author:  xingyu liao
+@contact: sherlockliao01@gmail.com
+"""
+import json
+import logging
+import os
+from collections import defaultdict
+import numpy as np
+import torch
+import torch.nn.functional as F
+from fastreid.evaluation import ReidEvaluator
+from fastreid.evaluation.query_expansion import aqe
+from fastreid.utils import comm
+from fastreid.utils.compute_dist import build_dist
+logger = logging.getLogger("fastreid.naic_submission")
+def partition_arg_topK(matrix, K, axis=0):
+    """
+    perform topK based on np.argpartition
+    :param matrix: to be sorted
+    :param K: select and sort the top K items
+    :param axis: 0 or 1. dimension to be sorted.
+    :return:
+    """
+    a_part = np.argpartition(matrix, K, axis=axis)
+    if axis == 0:
+        row_index = np.arange(matrix.shape[1 - axis])
+        a_sec_argsort_K = np.argsort(matrix[a_part[0:K, :], row_index], axis=axis)
+        return a_part[0:K, :][a_sec_argsort_K, row_index]
+    else:
+        column_index = np.arange(matrix.shape[1 - axis])[:, None]
+        a_sec_argsort_K = np.argsort(matrix[column_index, a_part[:, 0:K]], axis=axis)
+        return a_part[:, 0:K][column_index, a_sec_argsort_K]
+class NaicEvaluator(ReidEvaluator):
+    def process(self, inputs, outputs):
+        self.pids.extend(inputs["targets"])
+        self.camids.extend(inputs["camids"])
+        self.features.append(outputs.cpu())
+    def evaluate(self):
+        if comm.get_world_size() > 1:
+            comm.synchronize()
+            features = comm.gather(self.features)
+            features = sum(features, [])
+            pids = comm.gather(self.pids)
+            pids = sum(pids, [])
+            # fmt: off
+            if not comm.is_main_process(): return {}
+            # fmt: on
+        else:
+            features = self.features
+            pids = self.pids
+        features = torch.cat(features, dim=0)
+        # query feature, person ids and camera ids
+        query_features = features[:self._num_query]
+        query_pids = np.asarray(pids[:self._num_query])
+        # gallery features, person ids and camera ids
+        gallery_features = features[self._num_query:]
+        gallery_pids = np.asarray(pids[self._num_query:])
+        if self.cfg.TEST.AQE.ENABLED:
+            logger.info("Test with AQE setting")
+            qe_time = self.cfg.TEST.AQE.QE_TIME
+            qe_k = self.cfg.TEST.AQE.QE_K
+            alpha = self.cfg.TEST.AQE.ALPHA
+            query_features, gallery_features = aqe(query_features, gallery_features, qe_time, qe_k, alpha)
+        if self.cfg.TEST.METRIC == "cosine":
+            query_features = F.normalize(query_features, dim=1)
+            gallery_features = F.normalize(gallery_features, dim=1)
+        dist = build_dist(query_features, gallery_features, self.cfg.TEST.METRIC)
+        if self.cfg.TEST.RERANK.ENABLED:
+            logger.info("Test with rerank setting")
+            k1 = self.cfg.TEST.RERANK.K1
+            k2 = self.cfg.TEST.RERANK.K2
+            lambda_value = self.cfg.TEST.RERANK.LAMBDA
+            if self.cfg.TEST.METRIC == "cosine":
+                query_features = F.normalize(query_features, dim=1)
+                gallery_features = F.normalize(gallery_features, dim=1)
+            rerank_dist = build_dist(query_features, gallery_features, metric="jaccard", k1=k1, k2=k2)
+            dist = rerank_dist * (1 - lambda_value) + dist * lambda_value
+        if self.cfg.TEST.SAVE_DISTMAT:
+            np.save(os.path.join(self.cfg.OUTPUT_DIR, "distmat.npy"), dist)
+        results = defaultdict(list)
+        topk_indices = partition_arg_topK(dist, K=200, axis=1)[:, :200]
+        for i in range(topk_indices.shape[0]):
+            results[query_pids[i]].extend(gallery_pids[topk_indices[i]])
+        with open(os.path.join(self.cfg.OUTPUT_DIR, "submit.json"), 'w') as f:
+            json.dump(results, f)
+        return {}
--- a/projects/NAIC20/naic20r2_train_list_clean.txt
+++ b/projects/NAIC20/naic20r2_train_list_clean.txt
--- a/projects/NAIC20/train_list_clean.txt
+++ b/projects/NAIC20/train_list_clean.txt
--- a/projects/NAIC20/train_net.py
+++ b/projects/NAIC20/train_net.py
+# encoding: utf-8
+"""
+@author:  sherlock
+@contact: sherlockliao01@gmail.com
+"""
+import logging
+import sys
+sys.path.append('.')
+from fastreid.config import get_cfg
+from fastreid.engine import default_argument_parser, default_setup, launch
+from fastreid.utils.checkpoint import Checkpointer
+from fastreid.engine import DefaultTrainer
+from fastreid.data import build_reid_train_loader
+from naic import *
+class Trainer(DefaultTrainer):
+    @classmethod
+    def build_train_loader(cls, cfg):
+        logger = logging.getLogger("fastreid.naic20")
+        logger.info("Prepare NAIC20 competition trainset")
+        return build_reid_train_loader(cfg, rm_lt=cfg.DATASETS.RM_LT)
+class Committer(DefaultTrainer):
+    @classmethod
+    def build_evaluator(cls, cfg, dataset_name, output_dir=None):
+        data_loader, num_query = cls.build_test_loader(cfg, dataset_name)
+        return data_loader, NaicEvaluator(cfg, num_query, output_dir)
+def setup(args):
+    """
+    Create configs and perform basic setups.
+    """
+    cfg = get_cfg()
+    add_naic_config(cfg)
+    cfg.merge_from_file(args.config_file)
+    cfg.merge_from_list(args.opts)
+    cfg.freeze()
+    default_setup(cfg, args)
+    return cfg
+def main(args):
+    cfg = setup(args)
+    if args.eval_only:
+        cfg.defrost()
+        cfg.MODEL.BACKBONE.PRETRAIN = False
+        model = Trainer.build_model(cfg)
+        Checkpointer(model, save_dir=cfg.OUTPUT_DIR).load(cfg.MODEL.WEIGHTS)  # load trained model
+        if args.commit:
+            res = Committer.test(cfg, model)
+        else:
+            res = Trainer.test(cfg, model)
+        return res
+    trainer = Trainer(cfg)
+    trainer.resume_or_load(resume=args.resume)
+    return trainer.train()
+if __name__ == "__main__":
+    parser = default_argument_parser()
+    parser.add_argument("--commit", action="store_true", help="submission testing results")
+    args = parser.parse_args()
+    print("Command Line Args:", args)
+    launch(
+        main,
+        args.num_gpus,
+        num_machines=args.num_machines,
+        machine_rank=args.machine_rank,
+        dist_url=args.dist_url,
+        args=(args,),
+    )
--- a/projects/NAIC20/val_gallery.txt
+++ b/projects/NAIC20/val_gallery.txt
--- a/projects/NAIC20/val_query.txt
+++ b/projects/NAIC20/val_query.txt
+00051678.png:836
+00157560.png:4049
+00100903.png:8360
+00084036.png:1311
+00187556.png:19172
+00040480.png:15756
+00214678.png:14671
+00085683.png:13562
+00109425.png:15213
+00048168.png:6166
+00184092.png:20796
+00130870.png:23708
+00203859.png:21910
+00207172.png:14127
+00164065.png:22727
+00205570.png:7690
+00062169.png:20101
+00151257.png:9833
+00150266.png:23289
+00166594.png:9565
+00004491.png:6922
+00116069.png:7281
+00166204.png:14998
+00093527.png:28459
+00207549.png:13143
+00091601.png:1007
+00192020.png:20426
+00088494.png:15689
+00131046.png:1393
+00030322.png:25671
+00116134.png:1256
+00061779.png:20110
+00061152.png:7235
+00062448.png:9335
+00122921.png:12324
+00152228.png:466
+00114958.png:23941
+00170391.png:8635
+00108584.png:16104
+00124828.png:8271
+00196962.png:24426
+00044924.png:14026
+00170163.png:26132
+00155132.png:6125
+00068517.png:4027
+00224288.png:27880
+00225312.png:1906
+00007346.png:1104
+00058218.png:21696
+00190128.png:5413
+00031166.png:16736
+00225653.png:132
+00135673.png:17852
+00000247.png:813
+00038611.png:6902
+00183468.png:24672
+00169140.png:13759
+00028188.png:10001
+00172533.png:29114
+00046346.png:5203
+00202127.png:1230
+00116106.png:389
+00120823.png:5610
+00208986.png:10668
+00123495.png:23317
+00178645.png:17260
+00158312.png:24673
+00039291.png:14135
+00065190.png:11735
+00176841.png:14017
+00061768.png:7620
+00058944.png:8555
+00112234.png:22525
+00169918.png:29821
+00179783.png:5930
+00137025.png:21356
+00170948.png:14146
+00007968.png:8582
+00109230.png:15807
+00093274.png:3268
+00076315.png:11035
+00057003.png:14693
+00082290.png:17098
+00008836.png:17851
+00166975.png:9755
+00086380.png:103
+00159760.png:6708
+00078383.png:27037
+00034752.png:6557
+00039785.png:16220
+00058771.png:4997
+00111649.png:10121
+00031318.png:13186
+00026756.png:15289
+00008979.png:19766
+00144379.png:21424
+00062683.png:6594
+00218418.png:16683
+00040024.png:2353
+00012926.png:27171
+00139886.png:1649
+00101840.png:27964
+00024363.png:15098
+00039304.png:15637
+00096173.png:7275
+00179323.png:10833
+00103110.png:4355
+00042940.png:17075
+00006286.png:9767
+00154434.png:699
+00206386.png:15540
+00131566.png:16714
+00049103.png:10928
+00103687.png:6213
+00146667.png:9853
+00076072.png:26660
+00182565.png:14447
+00139942.png:10391
+00117899.png:5567
+00099911.png:20189
+00107260.png:14375
+00045411.png:27116
+00049284.png:23101
+00028170.png:14476
+00005544.png:14605
+00032714.png:19351
+00140124.png:4373
+00023111.png:24421
+00065138.png:14272
+00057613.png:9670
+00088121.png:13575
+00106926.png:18083
+00079299.png:11640
+00097234.png:29125
+00117911.png:16134
+00107750.png:9986
+00078454.png:8856
+00190655.png:22501
+00179116.png:24953
+00223239.png:4633
+00117114.png:28916
+00173903.png:14925
+00058976.png:27481
+00180937.png:10918
+00025843.png:11446
+00147730.png:10972
+00098839.png:28500
+00189548.png:25024
+00034194.png:3163
+00175309.png:17185
+00128349.png:5324
+00116515.png:2822
+00082081.png:25028
+00200063.png:13558
+00192986.png:26149
+00138834.png:12970
+00008001.png:23546
+00126738.png:29515
+00097632.png:6128
+00152332.png:11239
+00105725.png:27749
+00222564.png:27108
+00025718.png:19240
+00065519.png:12202
+00142768.png:28301
+00145367.png:19388
+00039702.png:12483
+00002463.png:13861
+00035298.png:24271
+00021509.png:11263
+00221756.png:23845
+00146505.png:23207
+00164879.png:3187
+00187948.png:19413
+00194876.png:25976
+00174819.png:7297
+00017468.png:18540
+00021808.png:14703
+00178170.png:10180
+00171695.png:20401
+00077090.png:21535
+00169763.png:1531
+00114557.png:21458
+00138183.png:20977
+00200809.png:6214
+00031873.png:17681
+00079710.png:8240
+00170535.png:804
+00076936.png:28854
+00147701.png:20805
+00221022.png:21057
+00054066.png:21682
+00072198.png:20357
+00086141.png:19779
+00186630.png:23213
+00145691.png:457
+00189514.png:19668
+00099511.png:7815
+00007382.png:17989
+00031664.png:12144
+00215254.png:12040
+00032527.png:7684
+00142579.png:28362
+00200113.png:12678
+00172859.png:15130
+00190530.png:3319
+00046185.png:6604
+00033681.png:6907
+00029152.png:28846
+00076713.png:5024
+00191441.png:1302
+00017141.png:18328
+00163283.png:23149
+00039288.png:25697
+00219010.png:20563
+00189010.png:24593
+00086475.png:7384
+00090727.png:16371
+00093383.png:26514
+00178385.png:19938
+00150507.png:21314
+00184076.png:26265
+00211950.png:24153
+00016435.png:28344
+00014075.png:14791
+00062976.png:2427
+00093974.png:26836
+00125902.png:18386
+00214956.png:9970
+00116019.png:23791
+00059512.png:27176
+00206831.png:10639
+00059015.png:21355
+00026562.png:18374
+00087530.png:23367
+00007783.png:12318
+00144887.png:12892
+00052712.png:11935
+00099024.png:25927
+00039551.png:384
+00080573.png:2245
+00106569.png:22248
+00076554.png:7828
+00216351.png:17269
+00092772.png:1708
+00171188.png:14239
+00157487.png:12129
+00030282.png:18365
+00059847.png:12922
+00058144.png:23787
+00153057.png:20527
+00076597.png:39
+00128140.png:21452
+00197562.png:13521
+00051311.png:6569
+00046794.png:20966
+00064850.png:3069
+00061546.png:18134
+00043187.png:16131
+00068248.png:3174
+00195883.png:12698
+00143058.png:17741
+00011243.png:28692
+00092603.png:11235
+00049805.png:17340
+00140209.png:1668
+00004049.png:14904
+00014210.png:2641
+00217276.png:7089
+00146916.png:24871
+00013476.png:5293
+00131844.png:2643
+00224851.png:2105
+00200119.png:18306
+00090020.png:1313
+00047350.png:10637
+00068989.png:529
+00046954.png:182
+00224121.png:28995
+00154763.png:2408
+00007211.png:26030
+00109565.png:5973
+00077165.png:2528
+00150327.png:16383
+00061091.png:13944
+00095710.png:5288
+00182465.png:23898
+00042544.png:4621
+00027249.png:4523
+00052768.png:9341
+00214021.png:10656
+00195757.png:2089
+00221063.png:5818
+00090129.png:5205
+00210235.png:27788
+00148923.png:27460
+00090037.png:29363
+00055710.png:27084
+00044710.png:6988
+00074022.png:16445
+00035507.png:3910
+00214769.png:27866
+00036115.png:20663
+00020244.png:12524
+00052502.png:10025
+00148412.png:3253
+00005297.png:14916
+00136777.png:13238
+00009700.png:7565
+00185673.png:10167
+00104840.png:7961
+00057646.png:185
+00035374.png:8645
+00038283.png:23536
+00130048.png:18223
+00221716.png:16673
+00177338.png:1426
+00154946.png:7654
+00058672.png:9506
+00033889.png:17486
+00216537.png:15639
+00139604.png:5016
+00088190.png:19442
+00212311.png:12068
+00055361.png:16540
+00207139.png:12341
+00075644.png:10846
+00189493.png:10717
+00056314.png:15560
+00084190.png:28290
+00198894.png:6571
+00048036.png:21862
+00041766.png:17025
+00056416.png:26848
+00147310.png:4862
+00103209.png:6896
+00093542.png:14445
+00204735.png:29727
+00152705.png:3438
+00123277.png:19955
+00185982.png:22562
+00149690.png:22241
+00083908.png:29522
+00067755.png:2693
+00001898.png:13053
+00086320.png:13069
+00065982.png:4888
+00215193.png:15824
+00180162.png:23499
+00112688.png:14454
+00074105.png:10473
+00178558.png:11595
+00219206.png:20193
+00156350.png:293
+00121220.png:11406
+00142465.png:16413
+00196000.png:19501
+00001179.png:28489
+00158850.png:26040
+00047528.png:25416
+00103502.png:16057
+00103577.png:4933
+00143502.png:7210
+00126713.png:24103
+00059634.png:13751
+00018645.png:9666
+00123300.png:2683
+00219827.png:21039
+00193770.png:3487
+00002301.png:1752
+00112593.png:22696
+00139069.png:13623
+00036318.png:12708
+00083178.png:2221
+00087193.png:20914
+00168390.png:4126
+00216841.png:13450
+00110769.png:17391
+00112648.png:24061
+00028117.png:21044
+00040105.png:26985
+00160576.png:10355
+00023760.png:16208
+00150475.png:22173
+00006442.png:10341
+00096819.png:21669
+00055322.png:26850
+00110732.png:23572
+00159031.png:9683
+00180888.png:17122
+00149365.png:967
+00071348.png:1998
+00066586.png:10178
+00008726.png:15935
+00217131.png:6450
+00194952.png:21564
+00066834.png:20430
+00032437.png:1193
+00188797.png:6933
+00065683.png:19221
+00001141.png:27367
+00002727.png:5498
+00163189.png:20848
+00218605.png:21263
+00175305.png:17879
+00027866.png:28829
+00188204.png:21093
+00173792.png:4082
+00082834.png:14900
+00102079.png:23372
+00217397.png:16789
+00192229.png:3907
+00088475.png:725
+00117160.png:27427
+00142395.png:24571
+00222154.png:10849
+00082787.png:12738
+00141645.png:7645
+00103964.png:250
+00206252.png:7080
+00001599.png:11932
+00080029.png:2332
+00038396.png:13378
+00168102.png:5043
+00139266.png:25943
+00085274.png:29472
+00174885.png:6765
+00101418.png:18121
+00088399.png:23283
+00063452.png:21655
+00200747.png:17322
+00086541.png:2689
+00010364.png:24528
+00108554.png:11902
+00193061.png:11828
+00068311.png:20476
+00225138.png:2356
+00054947.png:15388
+00206412.png:27997
+00050846.png:10826
+00109345.png:28702
+00067439.png:6853
+00208451.png:14793
+00025970.png:10129
+00131496.png:24605
+00125577.png:6244
+00067566.png:10435
+00094301.png:4448
+00108150.png:27035
+00130135.png:18052
+00179021.png:28317
+00124122.png:9938
+00046935.png:2830
+00121336.png:4930
+00108357.png:20943
+00113353.png:22494
+00044377.png:9699
+00128076.png:5030
+00200623.png:7833
+00182204.png:25705
+00140914.png:25339
+00169964.png:927
+00057058.png:12997
+00044958.png:25990
+00026979.png:4674
+00148014.png:10774
+00143937.png:27596
+00189334.png:23271
+00175429.png:3238
+00169301.png:17595
+00002361.png:20464
+00045101.png:22413
+00054092.png:21306
+00104423.png:22035
+00176956.png:24082
+00004002.png:11947
+00118020.png:1410
+00082426.png:18099
+00167382.png:21550
+00010486.png:976
+00097605.png:21042
+00161684.png:17725
+00091906.png:25789
+00018497.png:18123
+00220279.png:9865
+00066434.png:21563
+00023942.png:12854
+00025083.png:11405
+00150282.png:6580
+00087051.png:15324
+00155730.png:6110
+00098759.png:6100
+00170144.png:19961
+00028880.png:21326
+00186451.png:1492
+00154597.png:1213
+00189472.png:23901
+00053110.png:24262
+00212716.png:5071
+00159065.png:18122
+00157236.png:28666
+00183184.png:11573
+00177642.png:18673
+00188393.png:7835
+00224033.png:26537
+00048493.png:7781
+00069583.png:17057
+00122716.png:425
+00026586.png:27152
+00123744.png:24466
+00149062.png:18410
+00204465.png:2697
+00052062.png:25804
+00160512.png:26853
+00033688.png:13870
+00044946.png:6849
+00110177.png:20535
+00048286.png:26381
+00159406.png:16689
+00016870.png:10111
+00041611.png:13901
+00108195.png:14930
+00064164.png:13930
+00212079.png:23739
+00085349.png:28497
+00146517.png:26707
+00089396.png:23586
+00008308.png:29537
+00046799.png:7087
+00078952.png:15800
+00077366.png:8923
+00109850.png:15169
+00095959.png:6258
+00194365.png:27450
+00029959.png:12559
+00163915.png:3229
+00126141.png:16726
+00027134.png:28144
+00052361.png:10686
+00190014.png:4092
+00188110.png:21443
+00207900.png:13017
+00194901.png:13980
+00123003.png:20403
+00130616.png:12581
+00044712.png:16009
+00084741.png:24824
+00090113.png:4695
+00050957.png:22508
+00112902.png:19478
+00150045.png:12531
+00140734.png:5884
+00194855.png:24565
+00117260.png:28964
+00083065.png:11869
+00071093.png:2685
+00157242.png:21280
+00215640.png:25798
+00218852.png:29423
+00103572.png:29040
+00064882.png:7708
+00037001.png:8786
+00225040.png:885
+00213662.png:16899
+00024313.png:18109
+00048216.png:7834
+00147297.png:3983
+00091205.png:25283
+00072233.png:2244
+00105302.png:24023
+00068041.png:20574
+00099504.png:28743
+00203908.png:8429
+00156950.png:19628
+00169440.png:4198
+00032009.png:14225
+00140931.png:1543
+00015093.png:9916
+00144458.png:11759
+00137417.png:28416
+00209220.png:26006
+00004738.png:12930
+00171986.png:8767
+00079434.png:7628
+00050993.png:9561
+00217611.png:9901
+00198554.png:20016
+00052470.png:20134
+00217325.png:1938
+00062661.png:3877
+00026584.png:18424
+00072505.png:2504
+00135500.png:10343
+00142988.png:17561
+00087489.png:6061
+00032506.png:4868
+00059548.png:12412
+00103149.png:8536
+00152136.png:18576
+00125539.png:18104
+00001292.png:22624
+00025007.png:12383
+00003380.png:21319
+00142620.png:16945
+00053843.png:25194
+00199129.png:6705
+00030832.png:9613
+00186758.png:18492
+00091443.png:21139
+00099368.png:28350
+00010799.png:7066
+00053257.png:23121
+00017068.png:4387
+00179327.png:29505
+00110695.png:9578
+00024198.png:18675
+00128970.png:28725
+00172920.png:935
+00026048.png:25554
+00070562.png:4471
+00029878.png:18456
+00098765.png:29306
+00189315.png:27379
+00056636.png:316
+00115193.png:10139
+00074174.png:6454
+00129732.png:22291
+00140876.png:8722
+00171538.png:2104
+00038979.png:2094
+00208449.png:25217
+00132305.png:20675
+00003176.png:26946
+00125782.png:5020
+00119721.png:8069
+00169596.png:22137
+00121185.png:8056
+00189864.png:20099
+00089289.png:15901
+00062988.png:7635
+00132792.png:28160
+00102147.png:2619
+00033396.png:21795
+00110037.png:644
+00051095.png:3099
+00046103.png:24729
+00039767.png:13513
+00149306.png:20644
+00122332.png:14451
+00170817.png:6535
+00081398.png:34
+00179552.png:303
+00023395.png:21053
+00166532.png:24655
+00165315.png:21205
+00216356.png:10858
+00106690.png:29658
+00089776.png:20723
+00012766.png:3463
+00138710.png:4867
+00156769.png:21301
+00209061.png:18977
+00130807.png:24200
+00059498.png:29900
+00029387.png:13566
+00191829.png:24171
+00019887.png:150
+00110454.png:29415
+00106350.png:19618
+00088940.png:12647
+00195180.png:2484
+00035797.png:777
+00177289.png:15816
+00120093.png:3726
+00146385.png:23513
+00022450.png:27965
+00205233.png:28790
+00139841.png:22735
+00075327.png:29162
+00019179.png:26081
+00209023.png:11245
+00194429.png:25000
+00072933.png:8957
+00068151.png:8080
+00140855.png:8140
+00148185.png:10216
+00049640.png:13359
+00022741.png:28226
+00057038.png:23320
+00225736.png:11648
+00199526.png:21770
+00149024.png:12982
+00002696.png:2630
+00085614.png:15945
+00110486.png:6067
+00027778.png:27816
+00118497.png:16625
+00164495.png:21016
+00168112.png:28587
+00190561.png:16155
+00100477.png:27649
+00121495.png:17083
+00131181.png:11205
+00204898.png:19273
+00147445.png:26587
+00078479.png:25670
+00024514.png:19354
+00205452.png:27785
+00204827.png:9905
+00145419.png:21023
+00077473.png:10494
+00123567.png:842
+00148723.png:13868
+00177921.png:25195
+00173100.png:3854
+00088270.png:28872
+00066371.png:28823
+00095712.png:22074
+00057044.png:11276
+00190629.png:5826
+00191918.png:18801
+00180412.png:4602
+00037915.png:26319
+00063716.png:14245
+00070423.png:28322
+00174115.png:7024
+00033547.png:29596
+00118961.png:585
+00077726.png:15734
+00136579.png:24335
+00170342.png:8306
+00009265.png:15925
+00145807.png:17182
+00118486.png:27254
+00058030.png:14881
+00144047.png:1518
+00067631.png:27010
+00110066.png:6800
+00187965.png:11045
+00040621.png:27601
+00215001.png:27390
+00102031.png:13895
+00224627.png:21660
+00142961.png:22497
+00135887.png:21727
+00064687.png:18397
+00101354.png:29668
+00000514.png:15460
+00167615.png:8893
+00035958.png:4842
+00202959.png:2031
+00185114.png:1301
+00200032.png:26007
+00163235.png:27058
+00207621.png:10658
+00045329.png:11112
+00130729.png:27038
+00022031.png:25232
+00180783.png:618
+00079741.png:24120
+00060499.png:1276
+00153421.png:29436
+00104461.png:242
+00134461.png:23995
+00077148.png:27769
+00210757.png:19824
+00138026.png:25464
+00191380.png:5943
+00087769.png:16862
+00050715.png:26403
+00195378.png:24649
+00208518.png:16097
+00204256.png:3283
+00127954.png:8611
+00036608.png:15054
+00100494.png:21931
+00125336.png:7926
+00137371.png:15487
+00116041.png:13041
+00158550.png:24670
+00120863.png:12572
+00164441.png:15
+00117237.png:4071
+00081804.png:5932
+00209131.png:11066
+00108496.png:26885
+00194770.png:27336
+00025403.png:5415
+00025267.png:12196
+00105000.png:19856
+00068120.png:528
+00142347.png:16544
+00030265.png:6937
+00026379.png:10451
+00121066.png:16512
+00057461.png:21903
+00145277.png:12598
+00199030.png:14921
+00170677.png:7291
+00105442.png:4902
+00004903.png:25
+00109023.png:23373
+00202358.png:14444
+00040492.png:16081
+00157808.png:24256
+00114409.png:7295
+00091812.png:11147
+00056787.png:4716
+00024798.png:27678
+00016223.png:25289
+00189447.png:28326
+00003717.png:22128
+00163380.png:18973
+00139188.png:12974
+00200244.png:17160
+00096178.png:18166
+00023827.png:27301
+00224728.png:17109
+00161838.png:16569
+00086437.png:28643
+00065650.png:7531
+00196226.png:25567
+00034570.png:22628
+00119131.png:6595
+00107554.png:4242
+00016062.png:27501
+00143479.png:19670
+00050366.png:16687
+00085765.png:4262
+00035801.png:6980
+00175756.png:10674
+00079933.png:20529
+00058532.png:14757
+00143145.png:10291
+00064400.png:20733
+00079966.png:29078
+00000309.png:27989
+00141147.png:805
+00081011.png:6455
+00056168.png:8171
+00029068.png:19921
+00015840.png:29649
+00163062.png:22643
+00098349.png:6415
+00224703.png:18452
+00136832.png:2799
+00215537.png:17151
+00000552.png:9475
+00075837.png:1723
+00037226.png:14854
+00208678.png:1551
+00058578.png:12146
+00020207.png:2514
+00033222.png:6536
+00086533.png:3240
+00216199.png:25417
+00205230.png:9567
+00096945.png:2765
+00159982.png:24460
+00004146.png:10149
+00182794.png:10
+00172914.png:23481
+00058786.png:15060
+00010442.png:21179
+00124294.png:24316
+00147212.png:14646
+00192003.png:10450
+00135894.png:5619
+00023181.png:22150
+00130650.png:19031
+00152663.png:1550
+00059973.png:10461
+00210009.png:16675
+00184896.png:22294
+00144569.png:4212
+00107652.png:11754
+00114636.png:4153
+00204549.png:7107
+00069952.png:21880
+00131013.png:6136
+00122238.png:15288
+00104034.png:635
+00190710.png:21688
+00184001.png:17712
+00193616.png:7423
+00008103.png:10986
+00006281.png:18928
+00047533.png:26757
+00002521.png:14728
+00158129.png:17678
+00108766.png:28400
+00185823.png:21929
+00179952.png:19851
+00106359.png:10854
+00120010.png:26523
+00184729.png:18337
+00074599.png:18037
+00057998.png:739
+00180810.png:22846
+00075591.png:25991
+00015233.png:11231
+00216543.png:4645
+00207181.png:29483
+00204070.png:5605
+00126307.png:22107
+00046502.png:19585
+00032303.png:12562
+00016939.png:21900
+00002005.png:330
+00033453.png:22603
+00189378.png:2534
+00100866.png:8744
+00023249.png:25666
+00066923.png:29643
+00060966.png:19425
+00001208.png:12438
+00015808.png:20847
+00183947.png:12903
+00107014.png:760
+00086086.png:13553
+00158155.png:28640
+00008599.png:22547
+00047845.png:9523
+00102361.png:16977
+00083990.png:15602
+00095510.png:25669
+00016893.png:16441
+00071248.png:6117
+00094722.png:23789
+00135903.png:13854
+00199534.png:7213
+00111244.png:22543
+00112218.png:26823
+00120643.png:9443
+00144051.png:336
+00217001.png:9748
+00144270.png:25888
+00194982.png:3591
+00122370.png:4374
+00007702.png:3025
+00152071.png:12988
+00001401.png:5709
+00175233.png:15129
+00205109.png:28158
+00115570.png:21730
+00130987.png:3281
+00144387.png:29698
+00107913.png:4654
+00018501.png:23600
+00014409.png:24058
+00049022.png:16571
+00061071.png:24785
+00096509.png:19140
+00087408.png:14798
+00042581.png:5804
+00196495.png:18643
+00106966.png:14505
+00135946.png:11403
+00175430.png:8884
+00054299.png:14729
+00098047.png:12203
+00030851.png:15080
+00098762.png:27050
+00142979.png:7338
+00177589.png:9987
+00057167.png:27417
+00200311.png:23545
+00030044.png:26744
+00191575.png:3369
+00101357.png:20658
+00066221.png:7466
+00204248.png:13507
+00024165.png:29621
+00084784.png:20067
+00186610.png:18669
+00096866.png:7298
+00192203.png:10317
+00111386.png:490
+00054119.png:13241
+00198051.png:26763
+00138511.png:14824
+00069474.png:9521
+00148796.png:28775
+00150756.png:5054
+00083848.png:16027
+00194183.png:14562
+00124704.png:271
+00089333.png:5766
+00066567.png:9135
+00135928.png:5746
+00069889.png:26153
+00041920.png:29611
+00070870.png:9233
+00002225.png:24866
+00081953.png:24733
+00099554.png:11699
+00093216.png:11075
+00037281.png:14267
+00108215.png:5236
+00069076.png:15801
+00130337.png:11981
+00034858.png:15588
+00188534.png:28399
+00156689.png:1035
+00064506.png:7211
+00192545.png:8502
+00107202.png:3993
+00206407.png:26937
+00141094.png:29819
+00218489.png:4738
+00026108.png:21656
+00187138.png:9195
+00044664.png:1681
+00120301.png:28131
+00147674.png:3590
+00062269.png:4755
+00159954.png:14518
+00141256.png:5939
+00149329.png:8614
+00148606.png:7497
+00179899.png:19337
+00001222.png:26615
+00167214.png:16975
+00133501.png:17987
+00069509.png:26875
+00206660.png:29080
+00146202.png:12521
+00196126.png:19677
+00164428.png:25158
+00200852.png:9553
+00121906.png:27815
+00038322.png:4038
+00047650.png:22716
+00056817.png:1163
+00177180.png:29591
+00193338.png:1914
+00002029.png:18700
+00078392.png:2543
+00178620.png:10773
+00001249.png:16346
+00159639.png:5461
+00043938.png:24591
+00169988.png:17542
+00106336.png:2085
+00118570.png:5599
+00017163.png:11127
+00025878.png:20314
+00014505.png:27230
+00021931.png:5728
+00199676.png:1982
+00121778.png:1383
+00062029.png:25534
+00080996.png:26508
+00205314.png:18967
+00166505.png:6804
+00046668.png:14743
+00164710.png:11788
+00133803.png:8653
+00005403.png:27478
+00068610.png:17178
+00027788.png:19091
+00223067.png:9650
+00009324.png:19776
+00054567.png:4392
+00136207.png:16319
+00092707.png:25262
+00071808.png:20603
+00183601.png:9454
+00030904.png:4980
+00083302.png:13314
+00210772.png:8381
+00144580.png:20085
+00092550.png:24730
+00020844.png:1782
+00059550.png:28393
+00090781.png:9198
+00086053.png:6991
+00169064.png:6360
+00069507.png:9735
+00069939.png:26467
+00215604.png:13340
+00085697.png:23287
+00050198.png:25951
+00175724.png:22883
+00168462.png:26171
+00014799.png:14786
+00192811.png:1899
+00169823.png:12249
+00210096.png:26032
+00023109.png:12601
+00062487.png:21430
+00223704.png:8954
+00216501.png:8796
+00151916.png:19055
+00188392.png:12241
+00107399.png:1625
+00077006.png:13922
+00006670.png:7841
+00140310.png:29431
+00039410.png:12085
+00074504.png:29499
+00132313.png:18065
+00070255.png:12873
+00125235.png:26620
+00149921.png:10734
+00200042.png:21990
+00201410.png:19988
+00060720.png:22184
+00134606.png:25367
+00034374.png:16879
+00114373.png:5096
+00210709.png:1120
+00155425.png:8737
+00099323.png:24821
+00027650.png:3542
+00072452.png:29178
+00105903.png:13631
+00089132.png:27064
+00084412.png:16052
+00088894.png:14977
+00154440.png:3808
+00078254.png:16751
+00093203.png:10284
+00172394.png:16405
+00091119.png:6272
+00167990.png:3307
+00183802.png:5790
+00105660.png:19629
+00153464.png:9802
+00172137.png:13059
+00045309.png:13201
+00160397.png:971
+00119568.png:26360
+00111672.png:8930
+00028537.png:25311
+00169521.png:21601
+00183388.png:13239
+00187001.png:3491
+00052818.png:4805
+00089146.png:28523
+00226226.png:9981
+00091362.png:23903
+00033237.png:13140
+00217755.png:27694
+00163549.png:9538
+00026715.png:16519
+00225449.png:11120
+00223075.png:23135
+00218528.png:20825
+00103348.png:19937
+00137670.png:14118
+00070702.png:2930
+00205415.png:21009
+00180375.png:24692
+00137127.png:12879
+00135628.png:2944
+00213474.png:22582
+00133570.png:19701
+00153832.png:8473
+00031796.png:4788
+00007910.png:1283
+00074224.png:22876
+00188932.png:16886
+00213451.png:14924
+00168434.png:11744
+00151382.png:893
+00163009.png:11055
+00001088.png:20013
+00130594.png:16953
+00009035.png:4035
+00169926.png:4017
+00099918.png:22371
+00048267.png:1774
+00110014.png:10053
+00116940.png:27929
+00147848.png:22267
+00057740.png:12547
+00216114.png:7976
+00048669.png:24204
+00013204.png:14250
+00028396.png:18241
+00049911.png:15300
+00131667.png:913
+00112541.png:7824
+00170630.png:3681
+00202717.png:25672
+00040859.png:28549
+00178860.png:8305
+00121711.png:26459
+00055481.png:18120
+00156491.png:28396
+00126596.png:14529
+00160871.png:10532
+00068430.png:8715
+00128610.png:8629
+00215181.png:3791
+00034265.png:6886
+00150292.png:3615
+00148867.png:20812
+00170272.png:16196
+00188003.png:18339
+00063835.png:27337
+00135314.png:11861
+00085244.png:26960
+00068161.png:16920
+00179661.png:26412
+00043239.png:2434
+00201152.png:22819
+00036082.png:12010
+00166307.png:21684
+00146766.png:26096
+00012047.png:16686
+00214893.png:22392
+00204132.png:9621
+00208166.png:6792
+00133181.png:3441
+00011612.png:6833
+00205084.png:29894
+00222064.png:14613
+00093438.png:15464
+00118516.png:21030
+00009003.png:26168
+00157512.png:5175
+00000428.png:7353
+00061039.png:21317
+00204056.png:10932
+00171514.png:3085
+00144121.png:19160
+00062424.png:6484
+00130821.png:15682
+00206489.png:20042
+00161212.png:25353
+00223802.png:14293
+00089195.png:5159
+00160415.png:10981
+00095411.png:27402
+00051057.png:29103
+00031598.png:14347
+00039032.png:5505
+00139244.png:2990
+00216572.png:24319
+00082341.png:8813
+00196632.png:27363
+00208941.png:26313
+00083403.png:21772
+00023221.png:12793
+00095597.png:57
+00074882.png:17189
+00017899.png:4202
+00057904.png:6938
+00160673.png:8347
+00168566.png:18257
+00030738.png:21421
+00028866.png:13554
+00142972.png:96
+00145917.png:7871
+00195393.png:29049
+00176088.png:13027
+00005010.png:23185
+00112747.png:17292
+00106829.png:16830
+00181676.png:5792
+00053170.png:19461
+00097732.png:17117
+00081618.png:20456
+00034092.png:1033
+00179184.png:5347
+00180921.png:27942
+00108121.png:6863
+00003570.png:13885
+00024300.png:27202
+00007847.png:24198
+00223666.png:11940
+00089995.png:11934
+00204645.png:27583
+00203327.png:26788
+00216970.png:27410
+00176930.png:8144
+00001425.png:25079
+00015568.png:7827
+00002473.png:16992
+00194317.png:13808
+00130296.png:17101
+00214651.png:4488
+00127186.png:20662
+00080533.png:5233
+00059712.png:9533
+00127737.png:24898
+00111755.png:14034
+00208153.png:396
+00025551.png:11524
+00013048.png:25579
+00191007.png:18628
+00084480.png:19344
+00181871.png:11431
+00116819.png:26036
+00190431.png:28668
+00004207.png:9417
+00042782.png:460
+00048371.png:7109
+00062534.png:24393
+00096613.png:25745
+00139374.png:11088
+00111791.png:20579
+00002129.png:13322
+00161213.png:10728
+00203218.png:24494
+00200117.png:24293
+00206324.png:10193
+00170570.png:3668
+00042756.png:18390
+00071923.png:17332
+00119061.png:17815
+00212504.png:10258
+00159579.png:3444
+00176627.png:24607
+00182152.png:22320
+00005431.png:18859
+00014466.png:16073
+00041420.png:16607
+00114674.png:22016
+00123424.png:18874
+00008174.png:359
+00054243.png:22249
+00070888.png:3528
+00151863.png:8182
+00114314.png:6066
+00020426.png:1200
+00151849.png:21352
+00053881.png:25486
+00071521.png:342
+00184926.png:17131
+00178238.png:11785
+00027342.png:28966
+00038814.png:17480
+00095656.png:23194
+00052907.png:4659
+00172819.png:13593
+00028686.png:865
+00018507.png:10255
+00105836.png:10418
+00174726.png:21064
+00016103.png:19830
+00004698.png:23825
+00040651.png:4678
+00017410.png:10692
+00070663.png:21958
+00221507.png:21095
+00076421.png:16801
+00002071.png:9327
+00193069.png:16401
+00142905.png:9058
+00193439.png:1879
+00225366.png:21704
+00123316.png:14644
+00105132.png:9792
+00136853.png:756
+00081824.png:25892
+00020608.png:3150
+00041181.png:16192
+00079120.png:14574
+00022539.png:13283
+00168108.png:17465
+00176266.png:3101
+00205055.png:3703
+00173920.png:13957
+00103683.png:19446
+00184842.png:29771
+00022847.png:20882
+00210131.png:23907
+00221333.png:6603
+00181935.png:25448
+00170779.png:24872
+00048591.png:22488
+00128544.png:1436
+00161169.png:25906
+00061115.png:28064
+00016322.png:26797
+00040186.png:27995
+00009987.png:2961
+00083201.png:24671
+00012921.png:18474
+00142115.png:10845
+00058149.png:7634
+00071974.png:2533
+00143798.png:638
+00039736.png:15836
+00131110.png:22272
+00219947.png:1048
+00078098.png:11268
+00000893.png:26285
+00049685.png:9884
+00019151.png:26293
+00160455.png:18376
+00212620.png:590
+00101727.png:9668
+00194424.png:29578
+00109427.png:14902
+00193329.png:28845
+00020466.png:505
+00176990.png:9711
+00110684.png:3889
+00148591.png:11197
+00101084.png:26474
+00078271.png:11837
+00026111.png:27461
+00190455.png:20019
+00226142.png:13031
+00143881.png:1105
+00050357.png:21872
+00208657.png:24227
+00077957.png:11064
+00067112.png:21526
+00219696.png:19419
+00080224.png:10421
+00203603.png:26747
+00003913.png:295
+00173612.png:22613
+00079904.png:23387
+00031336.png:22474
+00010300.png:666
+00158946.png:17461
+00211508.png:23242
+00014860.png:9857
+00018605.png:24786
+00085959.png:21474
+00215556.png:6249
+00035689.png:3704
+00203919.png:22278
+00002422.png:27955
+00123939.png:13279
+00089565.png:27078
+00135988.png:25319
+00207062.png:1510
+00019286.png:18584
+00075990.png:24257
+00072272.png:28428
+00174522.png:25207
+00090625.png:17716
+00222492.png:3796
+00110042.png:25648
+00097784.png:1375
+00072460.png:25333
+00043373.png:3656
+00128879.png:27668
+00189212.png:16709
+00177574.png:28952
+00220198.png:16282
+00143562.png:10015
+00071861.png:22746
+00048805.png:2604
+00150052.png:28632
+00163550.png:15133
+00198165.png:1095
+00081122.png:22146
+00199616.png:12463
+00106230.png:795
+00108044.png:8985
+00206992.png:8836
+00135210.png:19672
+00141526.png:11023
+00024525.png:10431
+00207013.png:14350
+00138902.png:5704
+00046241.png:9194
+00085737.png:26140
+00070405.png:25104
+00066254.png:10174
+00124614.png:22265
+00127222.png:29464
+00122056.png:9651
+00156315.png:10172
+00024912.png:10610
+00047815.png:9956
+00056185.png:29007
+00057831.png:19874
+00169285.png:10074
+00093918.png:15341
+00079639.png:2574
+00143169.png:26445
+00124777.png:28203
+00017356.png:13981
+00192649.png:25054
+00175151.png:26177
+00185411.png:13210
+00041286.png:12386
+00103415.png:1671
+00146570.png:28046
+00037798.png:20201
+00052404.png:25561
+00205309.png:15482
+00050184.png:6728
+00038428.png:19843
+00207876.png:159
+00160482.png:21705
+00090424.png:3530
+00170665.png:24110
+00099853.png:28374
+00124647.png:24542
+00204950.png:10361
+00040382.png:1367
+00157640.png:6177
+00095721.png:5436
+00207987.png:4590
+00035245.png:14356
+00050153.png:22215
+00193391.png:23741
+00002835.png:6460
+00154129.png:11825
+00206132.png:19865
+00147759.png:720
+00072520.png:27745
+00115427.png:13546
+00082022.png:13771
+00044928.png:10381
+00123040.png:28323
+00047547.png:23649
+00069172.png:4486
+00015342.png:4335
+00002845.png:12788
+00003522.png:3115
+00173842.png:13867
+00167737.png:26039
+00085790.png:5455
+00134234.png:5226
+00195869.png:15123
+00138201.png:26201
+00222639.png:14413
+00110674.png:26786
+00033776.png:25358
+00078293.png:6088
+00108913.png:22243
+00141849.png:8142
+00218755.png:24025
+00054282.png:9379
+00184452.png:19297
+00208537.png:12379
+00162675.png:19873
+00017117.png:10417
+00221621.png:9014
+00088897.png:29153
+00087401.png:14231
+00004043.png:180
+00108165.png:6880
+00198246.png:7508
+00056112.png:13224
+00151048.png:15542
+00082527.png:11018
+00134399.png:6276
+00193518.png:16638
+00046717.png:4981
+00028742.png:8493
+00132442.png:24432
+00123400.png:12188
+00131511.png:13935
+00067837.png:17828
+00131917.png:20941
+00033978.png:29202
+00138199.png:17037
+00187349.png:22919
+00030086.png:10387
+00164767.png:1694
+00070494.png:20301
+00190286.png:1564
+00021964.png:28480
+00146717.png:2753
+00046468.png:13348
+00159341.png:1534
+00074814.png:12937
+00096686.png:5325
+00088652.png:14775
+00034893.png:16964
+00104402.png:21153
+00157045.png:25109
+00149206.png:16708
+00071778.png:19218
+00137553.png:8017
+00104360.png:3143
+00222321.png:2997
+00062785.png:6686
+00172790.png:29376
+00113541.png:3387
+00137067.png:21836
+00138323.png:23312
+00097490.png:6549
+00180189.png:12443
+00221159.png:11984
+00047852.png:2957
+00147106.png:2645
+00046327.png:27832
+00137242.png:12798
+00154079.png:18549
+00027389.png:26162
+00218219.png:28554
+00054394.png:9002
+00014861.png:8454
+00223087.png:11318
+00087169.png:14898
+00188765.png:12450
+00192917.png:14492
+00003773.png:17638
+00097936.png:15412
+00224690.png:6544
+00043153.png:16631
+00134907.png:50
+00195848.png:8038
+00028502.png:19149
+00219468.png:5768
+00193612.png:11793
+00078104.png:8065
+00202687.png:13797
+00219664.png:9709
+00212304.png:1573
+00152086.png:9069
+00031229.png:2788
+00200539.png:16724
+00071531.png:14655
+00176445.png:23868
+00178896.png:26799
+00041350.png:28875
+00088663.png:23752
+00165025.png:2871
+00217469.png:20728
+00111122.png:21606
+00017810.png:26535
+00185485.png:15338
+00034599.png:748
+00138386.png:6321
+00044641.png:24579
+00143712.png:21323
+00216431.png:18485
+00120686.png:13006
+00220639.png:23362
+00191669.png:29311
+00051529.png:23050
+00122998.png:12525
+00103820.png:26329
+00187275.png:1796
+00146669.png:26808
+00013662.png:27383
+00019294.png:8914
+00133617.png:20523
+00069756.png:692
+00195208.png:6225
+00083596.png:16918
+00204411.png:13170
+00042331.png:19340
+00197360.png:16711
+00135749.png:8798
+00141239.png:375
+00130317.png:14875
+00195581.png:10066
+00107585.png:12128
+00132912.png:9641
+00087281.png:23206
+00222254.png:21291
+00139111.png:21637
+00037047.png:3884
+00067652.png:6564
+00095761.png:21463
+00172026.png:9452
+00177931.png:21527
+00177162.png:20060
+00071502.png:28208
+00005654.png:18259
+00098215.png:29644
+00144348.png:17181
+00133036.png:21454
+00113022.png:5285
+00139844.png:22380
+00092793.png:19309
+00130195.png:18416
+00068343.png:14048
+00002725.png:28553
+00130793.png:12557
+00116166.png:21335
+00081390.png:9080
+00182158.png:3273
+00106539.png:18668
+00176992.png:29835
+00034526.png:16985
+00225310.png:11132
+00008962.png:7518
+00004965.png:3479
+00044485.png:15918
+00217923.png:29648
+00176092.png:196
+00114486.png:1413
+00175273.png:21592
+00211977.png:22364
+00113085.png:8583
+00084037.png:23019
+00021689.png:8280
+00079924.png:14248
+00147477.png:2346
+00168816.png:693
+00195566.png:24314
+00068592.png:3371
+00214111.png:11792
+00033349.png:27944
+00098290.png:19422
+00188171.png:6437
+00089815.png:27661
+00154555.png:4267
+00115390.png:27883
+00132407.png:3939
+00102154.png:21340
+00061618.png:27698
+00129773.png:14395
+00198437.png:5425
+00220336.png:13952
+00089382.png:13736
+00076953.png:26882
+00213753.png:24476
+00147433.png:23955
+00092278.png:4354
+00201304.png:18305
+00151832.png:9879
+00191604.png:20645
+00166300.png:5001
+00112929.png:7857
+00212543.png:15218
+00046440.png:21307
+00048118.png:3263
+00215358.png:5585
+00201676.png:20121
+00158673.png:25847
+00140550.png:24142
+00026630.png:29273
+00083565.png:7896
+00007057.png:20767
+00048637.png:13752
+00014291.png:11519
+00050214.png:18592
+00112914.png:21413
+00058223.png:23970
+00209068.png:1747
+00011622.png:22688
+00042134.png:21345
+00007993.png:28116
+00143861.png:10138
+00189011.png:13671
+00108011.png:2491
+00112608.png:28136
+00134784.png:22996
+00084496.png:2492
+00149622.png:1709
+00026094.png:13034
+00005798.png:7325
+00194254.png:16962
+00119493.png:19497
+00009404.png:11664
+00143595.png:13050
+00174277.png:26204
+00146405.png:6367
+00160586.png:25869
+00026254.png:8649
+00068575.png:3856
+00144779.png:10635
+00203868.png:26870
+00169225.png:9707
+00059502.png:29757
+00163871.png:29421
+00120826.png:6558
+00090159.png:28682
+00121669.png:21116
+00220741.png:23508
+00072856.png:8460
+00082815.png:1081
+00173783.png:21578
+00122255.png:89
+00214903.png:11436
+00109144.png:5551
+00030217.png:28405
+00147948.png:24690
+00008187.png:12596
+00048434.png:19725
+00167787.png:15630
+00010408.png:28237
+00143001.png:14812
+00203151.png:21668
+00120753.png:13942
+00129434.png:29262
+00062447.png:26645
+00130783.png:18852
+00094141.png:10882
+00079879.png:21265
+00216743.png:23620
+00065806.png:21598
+00141843.png:25468
+00049467.png:390
+00138761.png:28891
+00100869.png:20317
+00147696.png:2073
+00157911.png:25791
+00057655.png:7067
+00168420.png:26936
+00139270.png:4848
+00043267.png:2002
+00174847.png:4205
+00040541.png:24281
+00088248.png:29360
+00195779.png:8407
+00129129.png:11327
+00088200.png:3973
+00201497.png:28979
+00020889.png:16141
+00078330.png:22242
+00171401.png:28417
+00040001.png:19950
+00103534.png:28645
+00035925.png:4297
+00002868.png:4697
+00199460.png:5382
+00176201.png:23631
+00193185.png:23223
+00116452.png:6299
+00067638.png:15919
+00219891.png:23350
+00036251.png:25473
+00218814.png:4924
+00196704.png:25684
+00145783.png:11363
+00177320.png:14363
+00057994.png:25191
+00141851.png:12648
+00026372.png:25773
+00081623.png:24428
+00055455.png:17082
+00189595.png:13721
+00053089.png:9690
+00216665.png:8451
+00109561.png:4691
+00054777.png:11400
+00060190.png:23081
+00142924.png:176
+00139738.png:23696
+00031148.png:6365
+00032960.png:4165
+00220868.png:15833
+00105577.png:2537
+00211888.png:3157
+00099369.png:2580
+00071954.png:5050
+00099980.png:11679
+00178401.png:22732
+00080689.png:28732
+00057901.png:18530
+00011205.png:12871
+00151908.png:397
+00072678.png:16922
+00150412.png:18594
+00148339.png:14618
+00005254.png:22759
+00164356.png:6794
+00016005.png:4408
+00190676.png:24797
+00042824.png:13075
+00069403.png:12382
+00004013.png:18398
+00220483.png:8951
+00180886.png:8440
+00027435.png:10524
+00223637.png:15393
+00215694.png:10906
+00152271.png:947
+00087710.png:12002
+00165916.png:11102
+00058675.png:13446
+00174670.png:29823
+00176224.png:26012
+00016959.png:13005
+00072919.png:22844
+00007511.png:17225
+00015710.png:21769
+00104863.png:29726
+00213518.png:7293
+00093579.png:27542
+00003042.png:7638
+00123472.png:9842
+00045858.png:12593
+00162870.png:5477
+00150949.png:25177
+00213838.png:6864
+00158054.png:4841
+00167161.png:2365
+00033787.png:4722
+00063246.png:7106
+00158411.png:21119
+00211868.png:21566
+00191856.png:18093
+00213396.png:23984
+00112557.png:22580
+00078128.png:10303
+00138488.png:19997
+00041698.png:3308
+00118054.png:17762
+00148585.png:27107
+00021696.png:19477
+00124086.png:8487
+00142842.png:9442
+00049890.png:6583
+00222701.png:16191
+00166160.png:19593
+00146394.png:10311
+00066144.png:6388
+00035142.png:2906
+00139891.png:19979
+00123638.png:11360
+00028887.png:12413
+00061623.png:7991
+00056972.png:20152
+00150278.png:502
+00012186.png:5472
+00115293.png:8670
+00184167.png:153
+00048587.png:28877
+00057484.png:15962
+00091098.png:15884
+00182231.png:23837
+00020659.png:27884
+00020024.png:16312
+00204865.png:24634
+00121904.png:4290
+00199496.png:26829
+00133859.png:5964
+00198328.png:17010
+00211959.png:1878
+00146064.png:26567
+00166005.png:26345
+00133455.png:28697
+00200729.png:22130
+00118085.png:19453
+00128627.png:22933
+00124988.png:23643
+00105746.png:29848
+00102322.png:24114
+00148204.png:4762
+00125492.png:3335
+00014893.png:13946
+00029399.png:6957
+00031409.png:15159
+00200099.png:5721
+00071312.png:7826
+00113440.png:11996
+00116081.png:17857
+00190620.png:29027
+00217897.png:23410
+00121935.png:14043
+00027278.png:3915
+00036187.png:7237
+00181988.png:22434
+00136786.png:8201
+00173946.png:2124
+00050065.png:11715
+00060496.png:24366
+00146033.png:20487
+00128663.png:29046
+00183681.png:4832
+00090840.png:6395
+00199591.png:4863
+00016328.png:28027
+00027519.png:15095
+00200477.png:25063
+00080213.png:24202
+00044943.png:24799
+00203100.png:25682
+00013579.png:7185
+00188679.png:5431
+00224247.png:18091
+00180973.png:533
+00026146.png:26161
+00112690.png:16513
+00096628.png:18415
+00114186.png:13657
+00191035.png:12961
+00127028.png:27727
+00149354.png:14738
+00070849.png:13235
+00214969.png:12670
+00013120.png:4885
+00216369.png:11806
+00152507.png:839
+00218266.png:6824
+00054774.png:9438
+00095342.png:4149
+00127073.png:6340
+00101470.png:19926
+00112681.png:21147
+00024724.png:18706
+00075126.png:17584
+00138272.png:2756
+00164818.png:24184
+00173282.png:9862
+00102014.png:12589
+00011971.png:19742
+00130999.png:10709
+00076498.png:24389
+00065094.png:20659
+00192001.png:17276
+00168110.png:2523
+00025609.png:27616
+00182090.png:7498
+00018025.png:15890
+00056628.png:15234
+00032001.png:22815
+00135390.png:5776
+00179505.png:1807
+00046813.png:21744
+00124353.png:18991
+00180876.png:24517
+00200455.png:19064
+00046060.png:17235
+00190299.png:15969
+00016446.png:15651
+00109725.png:18716
+00094388.png:27143
+00152032.png:7982
+00190518.png:24584
+00023063.png:8361
+00090777.png:11124
+00169339.png:12756
+00023018.png:17517
+00079599.png:13203
+00010346.png:18146
+00006168.png:4882
+00127038.png:1825
+00132246.png:16565
+00100567.png:2520
+00199413.png:6639
+00014275.png:18163
+00086743.png:29294
+00163548.png:24038
+00142583.png:28172
+00089401.png:11450
+00095696.png:2992
+00151497.png:17768
+00023648.png:10497
+00141201.png:23653
+00174755.png:14182
+00105789.png:24029
+00076276.png:6866
+00118348.png:11016
+00199275.png:10869
+00189847.png:14207
+00043048.png:3556
+00012211.png:20465
+00045021.png:17242
+00101588.png:27898
+00008232.png:13686
+00058927.png:850
+00035390.png:22794
+00126911.png:5123
+00012137.png:16144
+00164831.png:26214
+00104921.png:5524
+00064441.png:23610
+00148369.png:25463
+00000151.png:16101
+00059175.png:8738
+00172431.png:22999
+00218663.png:23386
+00036743.png:493
+00206005.png:12418
+00055360.png:4636
+00213264.png:19581
+00034539.png:16298
+00145571.png:9875
+00212914.png:27979
+00080411.png:13019
+00087410.png:13700
+00128642.png:14837
+00207689.png:3054
+00117297.png:20294
+00043340.png:10841
+00095169.png:18743
+00110517.png:19279
+00077479.png:23291
+00071643.png:9652
+00217689.png:3984
+00208983.png:11141
+00006863.png:12796
+00012122.png:29455
+00225093.png:11604
+00191117.png:18487
+00032429.png:25989
+00211436.png:23912
+00186616.png:13579
+00154257.png:3179
+00141051.png:2115
+00140046.png:12498
+00193552.png:588
+00220118.png:20556
+00029926.png:22451
+00151972.png:7630
+00161566.png:29882
+00170394.png:18722
+00204975.png:21282
+00007291.png:24113
+00120684.png:1102
+00157398.png:25205
+00083068.png:14237
+00005918.png:20
+00129955.png:16154
+00072474.png:10913
+00186782.png:12567
+00194305.png:24754
+00117855.png:20959
+00076446.png:29378
+00182417.png:17264
+00081982.png:14958
+00114192.png:14202
+00209782.png:11634
+00067303.png:12183
+00115584.png:25125
+00077088.png:1323
+00086625.png:25858
+00206721.png:11852
+00044484.png:12300
+00063902.png:10719
+00139234.png:15188
+00189149.png:1947
+00051303.png:22783
+00105604.png:6688
+00090469.png:27933
+00132000.png:17023
+00087216.png:4104
+00174614.png:28622
+00215704.png:12471
+00205958.png:13983
+00180731.png:27734
+00155616.png:26957
+00040530.png:11209
+00003940.png:26613
+00146832.png:18624
+00053831.png:19371
+00158502.png:5815
+00129268.png:9283
+00118246.png:18429
+00067419.png:11857
+00178000.png:7786
+00008846.png:3806
+00210977.png:8807
+00104697.png:6942
+00217535.png:24134
+00103362.png:7471
+00119735.png:1928
+00194739.png:14607
+00005135.png:11283
+00028044.png:26150
+00115323.png:28356
+00220085.png:23254
+00086557.png:2325
+00023569.png:23853
+00194251.png:1134
+00104257.png:24107
+00166919.png:5518
+00211682.png:22190
+00097516.png:28169
+00225752.png:6146
+00093225.png:1015
+00141267.png:26094
+00001872.png:24965
+00102380.png:22064
+00139732.png:21556
+00191754.png:2749
+00193702.png:23237
+00102815.png:11795
+00090186.png:26045
+00115818.png:25014
+00157627.png:11602
+00072079.png:13049
+00079304.png:6862
+00026216.png:7111
+00086858.png:15743
+00002400.png:18516
+00106465.png:12286
+00205879.png:27732
+00030256.png:901
+00012135.png:17116
+00159950.png:11442
+00013654.png:20578
+00202080.png:10953
+00158598.png:4274
+00064602.png:5809
+00094559.png:9790
+00011201.png:2345
+00005277.png:22952
+00152056.png:28015
+00195621.png:21748
+00124070.png:4916
+00014010.png:18613
+00224174.png:25196
+00008162.png:25731
+00109852.png:16120
+00162881.png:10359
+00118001.png:253
+00104475.png:12950
+00195668.png:2849
+00146138.png:12923
+00022944.png:20824
+00103547.png:23621
+00100996.png:17299
+00009836.png:19229
+00224619.png:3207
+00060485.png:26452
+00117590.png:17538
+00104635.png:7952
+00007909.png:6649
+00145300.png:10804
+00088625.png:22704
+00142268.png:5154
+00028576.png:28752
+00133222.png:6038
+00152771.png:5188
+00061360.png:8750
+00203575.png:23633
+00144606.png:10611
+00081815.png:5386
+00183277.png:5516
+00087715.png:3349
+00182013.png:21006
+00197673.png:17843
+00045937.png:26941
+00206056.png:24552
+00136869.png:575
+00185278.png:14779
+00151099.png:262
+00217233.png:498
+00066877.png:27380
+00090560.png:4614
+00226143.png:14458
+00136645.png:12345
+00224681.png:2168
+00178574.png:17618
+00205686.png:22058
+00127015.png:19324
+00140109.png:21176
+00134680.png:7374
+00203894.png:25332
+00001497.png:29043
+00068298.png:5107
+00207873.png:17733
+00197428.png:12211
+00078314.png:25710
+00149555.png:25886
+00083686.png:9876
+00020027.png:22837
+00131689.png:24629
+00096513.png:23314
+00196733.png:10277
+00027230.png:19607
+00151988.png:2816
+00028230.png:780
+00148501.png:7899
+00219866.png:19394
+00049420.png:11110
+00067288.png:596
+00069605.png:11697
+00070862.png:12036
+00087153.png:2596
+00141653.png:1506
+00086680.png:27512
+00130581.png:8557
+00137047.png:24687
+00094880.png:22639
+00054485.png:10143
+00078151.png:23267
+00205159.png:29232
+00076722.png:24060
+00008203.png:26989
+00160146.png:25661
+00026973.png:29365
+00059134.png:11758
+00102729.png:17203
+00138874.png:1385
+00098711.png:13897
+00095963.png:29701
+00130982.png:20620
+00144785.png:3014
+00221213.png:2973
+00065552.png:28438
+00026594.png:26419
+00035681.png:22216
+00154463.png:14024
+00199040.png:6170
+00039929.png:14348
+00080749.png:13807
+00194271.png:21588
+00223997.png:3837
+00061762.png:26024
+00016095.png:15915
+00154029.png:5360
+00140730.png:9604
+00054760.png:24538
+00189799.png:15392
+00152556.png:14291
+00137382.png:29523
+00197008.png:7539
+00158303.png:21468
+00226119.png:122
+00068601.png:5877
+00073093.png:28749
+00021504.png:13732
+00070477.png:29305
+00150531.png:20686
+00036483.png:28980
+00134827.png:14016
+00062348.png:2774
+00049942.png:5422
+00113359.png:22495
+00052911.png:8376
+00114110.png:21266
+00082126.png:10769
+00092553.png:9380
+00030584.png:4068
+00139619.png:2860
+00051865.png:25539
+00183854.png:13350
+00070483.png:29485
+00032733.png:12304
+00058722.png:19435
+00072724.png:19541
+00199733.png:26298
+00110691.png:5839
+00036250.png:9356
+00132789.png:23702
+00158763.png:6799
+00151889.png:19529
+00179370.png:2763
+00034520.png:24383
+00146506.png:10455
+00213571.png:313
+00223890.png:12337
+00147473.png:15103
+00135553.png:16431
+00071021.png:18934
+00025936.png:8819
+00113283.png:5678
+00118715.png:19398
+00219676.png:9989
+00138980.png:3279
+00063323.png:1488
+00029866.png:19690
+00121070.png:20385
+00001530.png:17990
+00152186.png:8675
+00028210.png:5996
+00138729.png:26979
+00105456.png:9697
+00083404.png:13438
+00132762.png:10000
+00157077.png:17076
+00036657.png:18273
+00105527.png:18980
+00201063.png:21815
+00085760.png:13067
+00066814.png:10621
+00019098.png:16786
+00194209.png:27731
+00086355.png:2153
+00213934.png:27632
+00128745.png:23443
+00026559.png:3909
+00219009.png:12009
+00175121.png:21284
+00129900.png:29589
+00103329.png:27027
+00031672.png:11108
+00043401.png:25459
+00056159.png:26047
+00031998.png:10884
+00224434.png:1291
+00029059.png:23375
+00104187.png:28011
+00226197.png:5754
+00049887.png:4002
+00101722.png:1242
+00131685.png:26100
+00126267.png:13388
+00175937.png:2097
+00143867.png:1119
+00034090.png:16707
+00214015.png:14860
+00209400.png:4960
+00162460.png:18671
+00155458.png:23166
+00225021.png:12237
+00071618.png:16325
+00019569.png:8965
+00099380.png:13443
+00020462.png:23786
+00173029.png:24407
+00011043.png:24154
+00201530.png:21675
+00028874.png:4751
+00152901.png:6132
+00122993.png:17761
+00225495.png:17881
+00081931.png:5361
+00200047.png:9659
+00167535.png:17234
+00046269.png:12602
+00032420.png:14433
+00092636.png:10758
+00038442.png:14575
+00224914.png:6582
+00206520.png:26053
+00016772.png:1919
+00161660.png:11930
+00168408.png:17850
+00029991.png:21569
+00022406.png:90
+00179816.png:29385
+00148389.png:6692
+00223299.png:21723
+00081106.png:18047
+00009263.png:28008
+00132260.png:25122
+00199224.png:18368
+00035249.png:12440
+00156650.png:16215
+00068717.png:5214
+00147025.png:12714
+00033600.png:5003
+00113929.png:10397
+00056396.png:28302
+00086685.png:28503
+00168440.png:9059
+00108351.png:15022
+00213043.png:9269
+00121977.png:18633
+00020638.png:7820
+00191862.png:10574
+00219267.png:7091
+00142935.png:9608
+00051316.png:8545
+00056130.png:21681
+00084506.png:14446
+00191444.png:13088
+00070474.png:6676
+00168493.png:20976
+00000991.png:17201
+00085071.png:4564
+00112228.png:18278
+00203281.png:22927
+00048114.png:29752
+00129940.png:2890
+00174832.png:19409
+00031545.png:4059
+00225142.png:538
+00096689.png:17547
+00173270.png:23148
+00153046.png:2496
+00148990.png:1195
+00188816.png:21142
+00088656.png:19
+00097911.png:2661
+00217555.png:25302
+00019540.png:15755
+00123365.png:19499
+00200191.png:19281
+00144566.png:5991
+00005085.png:6671
+00147803.png:1227
+00133461.png:1929
+00111691.png:15696
+00031901.png:5021
+00135601.png:24833
+00167841.png:21201
+00221134.png:22177
+00028690.png:23466
+00098780.png:28505
+00045280.png:16451
+00202384.png:8046
+00135334.png:18838
+00133118.png:10379
+00009227.png:12691
+00169592.png:20167
+00020201.png:25376
+00151171.png:17068
+00052082.png:21171
+00114695.png:1599
+00214045.png:19169
+00138216.png:526
+00048630.png:28594
+00057162.png:311
+00183904.png:9378
+00133671.png:21968
+00049093.png:3105
+00160292.png:19214
+00096048.png:11352
+00023824.png:22852
+00042918.png:11916
+00013534.png:6064
+00003923.png:22489
+00026364.png:9954
+00144135.png:12526
+00027974.png:5023
+00213762.png:1352
+00034743.png:22405
+00078764.png:4984
+00030912.png:1046
+00222549.png:5410
+00211686.png:6139
+00154621.png:11599
+00098685.png:15778
+00140542.png:14428
+00219227.png:18085
+00146814.png:11006
+00207625.png:16396
+00009300.png:14947
+00217324.png:11224
+00122851.png:8399
+00057157.png:26693
+00204677.png:24506
+00111697.png:3733
+00032210.png:12260
+00031319.png:20173
+00180413.png:25145
+00138364.png:19269
+00148670.png:20519
+00190139.png:21089
+00201528.png:11206
+00187540.png:25709
+00182281.png:4401
+00072293.png:7123
+00076707.png:10900
+00117621.png:12295
+00188808.png:9844
+00140538.png:21240
+00059929.png:24884
+00037209.png:19811
+00017705.png:25263
+00211171.png:24443
+00059945.png:19700
+00086401.png:8118
+00044567.png:8129
+00217939.png:1394
+00186031.png:11034
+00093502.png:16614
+00118515.png:16046
+00147957.png:19175
+00190674.png:26835
+00052934.png:6749
+00112597.png:6422
+00089309.png:15989
+00087142.png:24342
+00005787.png:28581
+00221659.png:26778
+00169941.png:14767
+00027357.png:18898
+00188097.png:14871
+00195427.png:23103
+00125150.png:13177
+00134922.png:6301
+00154831.png:5051
+00132720.png:27398
+00105256.png:15882
+00203835.png:11453
+00034457.png:24807
+00139098.png:12023
+00061563.png:26943
+00074101.png:1695
+00120847.png:2762
+00128085.png:27653
+00154068.png:7874
+00132149.png:11598
+00167468.png:1505
+00133230.png:19878
+00129440.png:19734
+00036689.png:14886
+00209256.png:5824
+00171419.png:9124
+00162152.png:581
+00084447.png:6390
+00203282.png:26063
+00141590.png:4544
+00202437.png:14466
+00167286.png:12938
+00183445.png:18031
+00199831.png:23480
+00159253.png:20634
--- a/projects/PartialReID/README.md
+++ b/projects/PartialReID/README.md
+# DSR in FastReID
+**Deep Spatial Feature Reconstruction for Partial Person Re-identification**
+Lingxiao He, Xingyu Liao
+[[`CVPR2018`](http://openaccess.thecvf.com/content_cvpr_2018/papers/He_Deep_Spatial_Feature_CVPR_2018_paper.pdf)] [[`BibTeX`](#CitingDSR)] 
+**Foreground-aware Pyramid Reconstruction for Alignment-free Occluded Person Re-identification**
+Lingxiao He, Xingyu Liao
+[[`ICCV2019`](http://openaccess.thecvf.com/content_ICCV_2019/papers/He_Foreground-Aware_Pyramid_Reconstruction_for_Alignment-Free_Occluded_Person_Re-Identification_ICCV_2019_paper.pdf)] [[`BibTeX`](#CitingFPR)]
+## News！
+[1] The old_version code can be check in [old_version](https://github.com/JDAI-CV/Partial-Person-ReID), you can obtain the same result published in paper, and the new version code is updating, please waiting!
+## Installation
+First install FastReID, and then put Partial Datasets in directory datasets. The whole framework of FastReID-DSR is
+<div align="center">
+<img src="https://firebasestorage.googleapis.com/v0/b/firescript-577a2.appspot.com/o/imgs%2Fapp%2FSherlockWorkspace%2F1nVTE3Sn5c.jpg?alt=media&token=e7e9fcfc-4fc1-49c8-bcf4-c007028fdd25" width="700px" />
+</div>
+and the detail you can refer to
+## Datasets
+The datasets can find in [Google Drive](https://drive.google.com/file/d/1p7Jvo-RJhU_B6hf9eAhIEFNhvrzM5cdh/view?usp=sharing)
+PartialREID---gallery: 300 images of 60 ids, query: 300 images of 60 ids
+PartialiLIDS---gallery: 119 images of 119 ids, query: 119 images of 119 ids
+OccludedREID---gallery: 1,000 images of 200 ids, query: 1,000 images of 200 ids
+## Training and Evaluation
+To train a model, run:
+```bash
+python3 projects/PartialReID/train_net.py --config-file <config.yaml>
+```
+For example, to train the re-id network with IBN-ResNet-50 Backbone
+one should execute:
+```bash
+CUDA_VISIBLE_DEVICES='0,1,2,3' python3 projects/PartialReID/train_net.py --config-file 'projects/PartialReID/configs/partial_market.yml'
+```
+## Results
+| Method | PartialREID | OccludedREID | PartialiLIDS |
+|:--:|:--:|:--:|:--:|
+|   | Rank@1 (mAP)| Rank@1 (mAP)| Rank@1 (mAP)|
+| DSR (CVPR’18)  |73.7(68.1) |72.8(62.8)|64.3(58.1)| 
+| FPR (ICCV'19) | 81.0(76.6)|78.3(68.0)|68.1(61.8)| 
+| FastReID-DSR | 82.7(76.8)|81.6(70.9)|73.1(79.8) | 
+## <a name="CitingDSR"></a >Citing DSR and Citing FPR
+If you use DSR or FPR, please use the following BibTeX entry.
+```
+@inproceedings{he2018deep,
+  title={Deep spatial feature reconstruction for partial person re-identification: Alignment-free approach},
+  author={He, Lingxiao and Liang, Jian and Li, Haiqing and Sun, Zhenan},
+  booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
+  year={2018}
+}
+@inproceedings{he2019foreground,
+  title={Foreground-aware Pyramid Reconstruction for Alignment-free Occluded Person Re-identification},
+  author={He, Lingxiao and Wang, Yinggang and Liu, Wu and Zhao, He and Sun, Zhenan and Feng, Jiashi},
+  booktitle={IEEE International Conference on Computer Vision (ICCV)},
+  year={2019}
+}
+```
--- a/projects/PartialReID/configs/partial_market.yml
+++ b/projects/PartialReID/configs/partial_market.yml
+MODEL:
+  META_ARCHITECTURE: PartialBaseline
+  BACKBONE:
+    NAME: build_resnet_backbone
+    NORM: BN
+    DEPTH: 50x
+    LAST_STRIDE: 1
+    FEAT_DIM: 2048
+    WITH_IBN: True
+    PRETRAIN: True
+  HEADS:
+    NAME: DSRHead
+    POOL_LAYER: FastGlobalAvgPool
+    WITH_BNNECK: True
+    CLS_LAYER: Linear
+  LOSSES:
+    NAME: ("CrossEntropyLoss", "TripletLoss",)
+    CE:
+      EPSILON: 0.12
+      SCALE: 1.
+    TRI:
+      MARGIN: 0.3
+      SCALE: 1.0
+      HARD_MINING: False
+DATASETS:
+  NAMES: ("Market1501",)
+  TESTS: ("PartialREID", "PartialiLIDS", "OccludedREID",)
+INPUT:
+  SIZE_TRAIN: [384, 128]
+  SIZE_TEST: [384, 128]
+  FLIP:
+    ENABLED: True
+DATALOADER:
+  SAMPLER_TRAIN: NaiveIdentitySampler
+  NUM_INSTANCE: 4
+  NUM_WORKERS: 8
+SOLVER:
+  AMP:
+    ENABLED: False
+  OPT: Adam
+  MAX_EPOCH: 60
+  BASE_LR: 0.0007
+  BIAS_LR_FACTOR: 1.
+  WEIGHT_DECAY: 0.0005
+  WEIGHT_DECAY_BIAS: 0.0005
+  IMS_PER_BATCH: 256
+  SCHED: CosineAnnealingLR
+  DELAY_EPOCHS: 20
+  ETA_MIN_LR: 0.0000007
+  WARMUP_FACTOR: 0.1
+  WARMUP_ITERS: 500
+  CHECKPOINT_PERIOD: 20
+TEST:
+  EVAL_PERIOD: 10
+  IMS_PER_BATCH: 128
+CUDNN_BENCHMARK: True
+OUTPUT_DIR: projects/PartialReID/logs/test_partial
\ No newline at end of file
--- a/projects/PartialReID/partialreid/__init__.py
+++ b/projects/PartialReID/partialreid/__init__.py
+# encoding: utf-8
+"""
+@author:  xingyu liao
+@contact: sherlockliao01@gmail.com
+"""
+from .partial_dataset import *
+from .partialbaseline import PartialBaseline
+from .dsr_head import DSRHead
+from .config import add_partialreid_config
+from .dsr_evaluation import DsrEvaluator
--- a/projects/PartialReID/partialreid/config.py
+++ b/projects/PartialReID/partialreid/config.py
+# encoding: utf-8
+"""
+@author:  l1aoxingyu
+@contact: sherlockliao01@gmail.com
+"""
+from fastreid.config import CfgNode as CN
+def add_partialreid_config(cfg):
+    _C = cfg
+    _C.TEST.DSR = CN({"ENABLED": True})
--- a/projects/PartialReID/partialreid/dsr_distance.py
+++ b/projects/PartialReID/partialreid/dsr_distance.py
+"""Numpy version of euclidean distance, etc.
+Notice the input/output shape of methods, so that you can better understand
+the meaning of these methods."""
+import numpy as np
+import torch
+def normalize(nparray, order=2, axis=0):
+    """Normalize a N-D numpy array along the specified axis."""
+    norm = np.linalg.norm(nparray, ord=order, axis=axis, keepdims=True)
+    return nparray / (norm + np.finfo(np.float32).eps)
+def compute_dsr_dist(array1, array2, distmat, scores):
+    """ Compute the sptial feature reconstruction of all pairs
+     array: [M, N, C] M: the number of query, N: the number of spatial feature, C: the dimension of each spatial feature
+     array2: [M, N, C] M: the number of gallery
+    :return:
+    numpy array with shape [m1, m2]
+    """
+    dist = 100 * torch.ones(len(array1), len(array2))
+    dist = dist.cuda()
+    kappa = 0.001
+    index = np.argsort(distmat, axis=1)
+    T = kappa * torch.eye(110)
+    T = T.cuda()
+    M = []
+    for i in range(0, len(array2)):
+        g = array2[i]
+        g = torch.FloatTensor(g)
+        g = g.view(g.size(0), g.size(1))
+        g = g.cuda()
+        Proj_M1 = torch.matmul(torch.inverse(torch.matmul(g.t(), g) + T), g.t())
+        Proj_M1 = Proj_M1.cpu().numpy()
+        M.append(Proj_M1)
+    for i in range(0, len(array1)):
+        q = torch.FloatTensor(array1[i])
+        q = q.view(q.size(0), q.size(1))
+        q = q.cuda()
+        for j in range(0, 100):
+            g = array2[index[i, j]]
+            g = torch.FloatTensor(g)
+            g = g.view(g.size(0), g.size(1))
+            g = g.cuda()
+            Proj_M = torch.FloatTensor(M[index[i, j]])
+            Proj_M = Proj_M.cuda()
+            a = torch.matmul(g, torch.matmul(Proj_M, q)) - q
+            dist[i, index[i, j]] = ((torch.pow(a, 2).sum(0).sqrt()) * scores[i].cuda()).sum()
+    dist = dist.cpu()
+    dist = dist.numpy()
+    return dist
--- a/projects/PartialReID/partialreid/dsr_evaluation.py
+++ b/projects/PartialReID/partialreid/dsr_evaluation.py
+# encoding: utf-8
+"""
+@author:  liaoxingyu
+@contact: sherlockliao01@gmail.com
+"""
+import copy
+import logging
+from collections import OrderedDict
+import numpy as np
+import torch
+import torch.nn.functional as F
+from fastreid.evaluation.evaluator import DatasetEvaluator
+from fastreid.evaluation.rank import evaluate_rank
+from fastreid.utils import comm
+from .dsr_distance import compute_dsr_dist
+logger = logging.getLogger('fastreid.partialreid.dsr_evaluation')
+class DsrEvaluator(DatasetEvaluator):
+    def __init__(self, cfg, num_query, output_dir=None):
+        self.cfg = cfg
+        self._num_query = num_query
+        self._output_dir = output_dir
+        self.features = []
+        self.spatial_features = []
+        self.scores = []
+        self.pids = []
+        self.camids = []
+    def reset(self):
+        self.features = []
+        self.spatial_features = []
+        self.scores = []
+        self.pids = []
+        self.camids = []
+    def process(self, inputs, outputs):
+        self.pids.extend(inputs["targets"])
+        self.camids.extend(inputs["camids"])
+        self.features.append(F.normalize(outputs[0]).cpu())
+        outputs1 = F.normalize(outputs[1].data).cpu()
+        self.spatial_features.append(outputs1)
+        self.scores.append(outputs[2].cpu())
+    def evaluate(self):
+        if comm.get_world_size() > 1:
+            comm.synchronize()
+            features = comm.gather(self.features)
+            features = sum(features, [])
+            spatial_features = comm.gather(self.spatial_features)
+            spatial_features = sum(spatial_features, [])
+            scores = comm.gather(self.scores)
+            scores = sum(scores, [])
+            pids = comm.gather(self.pids)
+            pids = sum(pids, [])
+            camids = comm.gather(self.camids)
+            camids = sum(camids, [])
+            # fmt: off
+            if not comm.is_main_process(): return {}
+            # fmt: on
+        else:
+            features = self.features
+            spatial_features = self.spatial_features
+            scores = self.scores
+            pids = self.pids
+            camids = self.camids
+        features = torch.cat(features, dim=0)
+        spatial_features = torch.cat(spatial_features, dim=0).numpy()
+        scores = torch.cat(scores, dim=0)
+        # query feature, person ids and camera ids
+        query_features = features[:self._num_query]
+        query_pids = np.asarray(pids[:self._num_query])
+        query_camids = np.asarray(camids[:self._num_query])
+        # gallery features, person ids and camera ids
+        gallery_features = features[self._num_query:]
+        gallery_pids = np.asarray(pids[self._num_query:])
+        gallery_camids = np.asarray(camids[self._num_query:])
+        if self.cfg.TEST.METRIC == "cosine":
+            query_features = F.normalize(query_features, dim=1)
+            gallery_features = F.normalize(gallery_features, dim=1)
+        dist = 1 - torch.mm(query_features, gallery_features.t()).numpy()
+        self._results = OrderedDict()
+        query_features = query_features.numpy()
+        gallery_features = gallery_features.numpy()
+        if self.cfg.TEST.DSR.ENABLED:
+            logger.info("Testing with DSR setting")
+            dsr_dist = compute_dsr_dist(spatial_features[:self._num_query], spatial_features[self._num_query:], dist,
+                                        scores[:self._num_query])
+            max_value = 0
+            k = 0
+            for i in range(0, 101):
+                lamb = 0.01 * i
+                dist1 = (1 - lamb) * dist + lamb * dsr_dist
+                cmc, all_AP, all_INP = evaluate_rank(dist1, query_pids, gallery_pids, query_camids, gallery_camids)
+                if (cmc[0] > max_value):
+                    k = lamb
+                    max_value = cmc[0]
+            dist1 = (1 - k) * dist + k * dsr_dist
+            cmc, all_AP, all_INP = evaluate_rank(dist1, query_pids, gallery_pids, query_camids, gallery_camids)
+        else:
+            cmc, all_AP, all_INP = evaluate_rank(dist, query_pids, gallery_pids, query_camids, gallery_camids)
+        mAP = np.mean(all_AP)
+        mINP = np.mean(all_INP)
+        for r in [1, 5, 10]:
+            self._results['Rank-{}'.format(r)] = cmc[r - 1] * 100
+        self._results['mAP'] = mAP * 100
+        self._results['mINP'] = mINP * 100
+        return copy.deepcopy(self._results)
--- a/projects/PartialReID/partialreid/dsr_head.py
+++ b/projects/PartialReID/partialreid/dsr_head.py
+# encoding: utf-8
+"""
+@author:  lingxiao he
+@contact: helingxiao3@jd.com
+"""
+import torch
+import torch.nn.functional as F
+from torch import nn
+from fastreid.layers import *
+from fastreid.modeling.heads import EmbeddingHead
+from fastreid.modeling.heads.build import REID_HEADS_REGISTRY
+from fastreid.layers.weight_init import weights_init_kaiming
+class OcclusionUnit(nn.Module):
+    def __init__(self, in_planes=2048):
+        super(OcclusionUnit, self).__init__()
+        self.MaxPool1 = nn.MaxPool2d(kernel_size=2, stride=2, padding=0)
+        self.MaxPool2 = nn.MaxPool2d(kernel_size=4, stride=2, padding=0)
+        self.MaxPool3 = nn.MaxPool2d(kernel_size=6, stride=2, padding=0)
+        self.MaxPool4 = nn.MaxPool2d(kernel_size=8, stride=2, padding=0)
+        self.mask_layer = nn.Linear(in_planes, 1, bias=True)
+    def forward(self, x):
+        SpaFeat1 = self.MaxPool1(x)  # shape: [n, c, h, w]
+        SpaFeat2 = self.MaxPool2(x)
+        SpaFeat3 = self.MaxPool3(x)
+        SpaFeat4 = self.MaxPool4(x)
+        Feat1 = SpaFeat1.view(SpaFeat1.size(0), SpaFeat1.size(1), SpaFeat1.size(2) * SpaFeat1.size(3))
+        Feat2 = SpaFeat2.view(SpaFeat2.size(0), SpaFeat2.size(1), SpaFeat2.size(2) * SpaFeat2.size(3))
+        Feat3 = SpaFeat3.view(SpaFeat3.size(0), SpaFeat3.size(1), SpaFeat3.size(2) * SpaFeat3.size(3))
+        Feat4 = SpaFeat4.view(SpaFeat4.size(0), SpaFeat4.size(1), SpaFeat4.size(2) * SpaFeat4.size(3))
+        SpatialFeatAll = torch.cat((Feat1, Feat2, Feat3, Feat4), 2)
+        SpatialFeatAll = SpatialFeatAll.transpose(1, 2)  # shape: [n, c, m]
+        y = self.mask_layer(SpatialFeatAll)
+        mask_weight = torch.sigmoid(y[:, :, 0])
+        # mask_score = torch.sigmoid(mask_weight[:, :48])
+        feat_dim = SpaFeat1.size(2) * SpaFeat1.size(3)
+        mask_score = F.normalize(mask_weight[:, :feat_dim], p=1, dim=1)
+        #       mask_score_norm = mask_score
+        # mask_weight_norm = torch.sigmoid(mask_weight)
+        mask_weight_norm = F.normalize(mask_weight, p=1, dim=1)
+        mask_score = mask_score.unsqueeze(1)
+        SpaFeat1 = SpaFeat1.transpose(1, 2)
+        SpaFeat1 = SpaFeat1.transpose(2, 3)  # shape: [n, h, w, c]
+        SpaFeat1 = SpaFeat1.view((SpaFeat1.size(0), SpaFeat1.size(1) * SpaFeat1.size(2), -1))  # shape: [n, h*w, c]
+        global_feats = mask_score.matmul(SpaFeat1).view(SpaFeat1.shape[0], -1, 1, 1)
+        return global_feats, mask_weight, mask_weight_norm
+class Flatten(nn.Module):
+    def forward(self, input):
+        return input.view(input.size(0), -1)
+@REID_HEADS_REGISTRY.register()
+class DSRHead(EmbeddingHead):
+    def __init__(self, cfg):
+        super().__init__(cfg)
+        feat_dim = cfg.MODEL.BACKBONE.FEAT_DIM
+        with_bnneck = cfg.MODEL.HEADS.WITH_BNNECK
+        norm_type = cfg.MODEL.HEADS.NORM
+        num_classes = cfg.MODEL.HEADS.NUM_CLASSES
+        embedding_dim = cfg.MODEL.HEADS.EMBEDDING_DIM
+        self.occ_unit = OcclusionUnit(in_planes=feat_dim)
+        self.MaxPool1 = nn.MaxPool2d(kernel_size=2, stride=2, padding=0)
+        self.MaxPool2 = nn.MaxPool2d(kernel_size=4, stride=2, padding=0)
+        self.MaxPool3 = nn.MaxPool2d(kernel_size=6, stride=2, padding=0)
+        self.MaxPool4 = nn.MaxPool2d(kernel_size=8, stride=2, padding=0)
+        occ_neck = []
+        if embedding_dim > 0:
+            occ_neck.append(nn.Conv2d(feat_dim, embedding_dim, 1, 1, bias=False))
+            feat_dim = embedding_dim
+        if with_bnneck:
+            occ_neck.append(get_norm(norm_type, feat_dim, bias_freeze=True))
+        self.bnneck_occ = nn.Sequential(*occ_neck)
+        self.bnneck_occ.apply(weights_init_kaiming)
+        self.weight_occ = nn.Parameter(torch.normal(0, 0.01, (num_classes, feat_dim)))
+    def forward(self, features, targets=None):
+        """
+        See :class:`ReIDHeads.forward`.
+        """
+        SpaFeat1 = self.MaxPool1(features)  # shape: [n, c, h, w]
+        SpaFeat2 = self.MaxPool2(features)
+        SpaFeat3 = self.MaxPool3(features)
+        SpaFeat4 = self.MaxPool4(features)
+        Feat1 = SpaFeat1.view(SpaFeat1.size(0), SpaFeat1.size(1), SpaFeat1.size(2) * SpaFeat1.size(3))
+        Feat2 = SpaFeat2.view(SpaFeat2.size(0), SpaFeat2.size(1), SpaFeat2.size(2) * SpaFeat2.size(3))
+        Feat3 = SpaFeat3.view(SpaFeat3.size(0), SpaFeat3.size(1), SpaFeat3.size(2) * SpaFeat3.size(3))
+        Feat4 = SpaFeat4.view(SpaFeat4.size(0), SpaFeat4.size(1), SpaFeat4.size(2) * SpaFeat4.size(3))
+        SpatialFeatAll = torch.cat((Feat1, Feat2, Feat3, Feat4), dim=2)
+        foreground_feat, mask_weight, mask_weight_norm = self.occ_unit(features)
+        # print(time.time() - st)
+        bn_foreground_feat = self.bnneck_occ(foreground_feat)
+        bn_foreground_feat = bn_foreground_feat[..., 0, 0]
+        # Evaluation
+        if not self.training:
+            return bn_foreground_feat, SpatialFeatAll, mask_weight_norm
+        # Training
+        global_feat = self.pool_layer(features)
+        bn_feat = self.bottleneck(global_feat)
+        bn_feat = bn_feat[..., 0, 0]
+        if self.cls_layer.__class__.__name__ == 'Linear':
+            pred_class_logits = F.linear(bn_feat, self.weight)
+            fore_pred_class_logits = F.linear(bn_foreground_feat, self.weight_occ)
+        else:
+            pred_class_logits = F.linear(F.normalize(bn_feat), F.normalize(self.weight))
+            fore_pred_class_logits = F.linear(F.normalize(bn_foreground_feat), F.normalize(self.weight_occ))
+        cls_outputs = self.cls_layer(pred_class_logits, targets)
+        fore_cls_outputs = self.cls_layer(fore_pred_class_logits, targets)
+        # pdb.set_trace()
+        return {
+            "cls_outputs": cls_outputs,
+            "fore_cls_outputs": fore_cls_outputs,
+            "pred_class_logits": pred_class_logits * self.cls_layer.s,
+            "features": global_feat[..., 0, 0],
+            "foreground_features": foreground_feat[..., 0, 0],
+        }
--- a/projects/PartialReID/partialreid/partial_dataset.py
+++ b/projects/PartialReID/partialreid/partial_dataset.py
+# encoding: utf-8
+"""
+@author:  lingxiao he
+@contact: helingxiao3@jd.com
+"""
+import glob
+import os
+import os.path as osp
+import re
+from fastreid.data.datasets import DATASET_REGISTRY
+from fastreid.data.datasets.bases import ImageDataset
+__all__ = ['PartialREID', 'PartialiLIDS', 'OccludedREID']
+def process_test(query_path, gallery_path):
+    query_img_paths = glob.glob(os.path.join(query_path, '*.jpg'))
+    gallery_img_paths = glob.glob(os.path.join(gallery_path, '*.jpg'))
+    query_paths = []
+    pattern = re.compile(r'([-\d]+)_(\d*)')
+    for img_path in query_img_paths:
+        pid, camid = map(int, pattern.search(img_path).groups())
+        query_paths.append([img_path, pid, camid])
+    gallery_paths = []
+    for img_path in gallery_img_paths:
+        pid, camid = map(int, pattern.search(img_path).groups())
+        gallery_paths.append([img_path, pid, camid])
+    return query_paths, gallery_paths
+@DATASET_REGISTRY.register()
+class PartialREID(ImageDataset):
+    dataset_name = "partialreid"
+    def __init__(self, root='datasets',):
+        self.root = root
+        self.query_dir = osp.join(self.root, 'Partial_REID/partial_body_images')
+        self.gallery_dir = osp.join(self.root, 'Partial_REID/whole_body_images')
+        query, gallery = process_test(self.query_dir, self.gallery_dir)
+        ImageDataset.__init__(self, [], query, gallery)
+@DATASET_REGISTRY.register()
+class PartialiLIDS(ImageDataset):
+    dataset_name = "partialilids"
+    def __init__(self, root='datasets',):
+        self.root = root
+        self.query_dir = osp.join(self.root, 'PartialiLIDS/query')
+        self.gallery_dir = osp.join(self.root, 'PartialiLIDS/gallery')
+        query, gallery = process_test(self.query_dir, self.gallery_dir)
+        ImageDataset.__init__(self, [], query, gallery)
+@DATASET_REGISTRY.register()
+class OccludedREID(ImageDataset):
+    dataset_name = "occludereid"
+    def __init__(self, root='datasets',):
+        self.root = root
+        self.query_dir = osp.join(self.root, 'OccludedREID/query')
+        self.gallery_dir = osp.join(self.root, 'OccludedREID/gallery')
+        query, gallery = process_test(self.query_dir, self.gallery_dir)
+        ImageDataset.__init__(self, [], query, gallery)