| 96 | |
| 97 | |
| 98 | class RefDataset(Dataset): |
| 99 | def __init__(self, lmdb_dir, mask_dir, dataset, split, mode, input_size, |
| 100 | word_length): |
| 101 | super(RefDataset, self).__init__() |
| 102 | self.lmdb_dir = lmdb_dir |
| 103 | self.mask_dir = mask_dir |
| 104 | self.dataset = dataset |
| 105 | self.split = split |
| 106 | self.mode = mode |
| 107 | self.input_size = (input_size, input_size) |
| 108 | #self.mask_size = [13, 26, 52] |
| 109 | self.word_length = word_length |
| 110 | self.mean = torch.tensor([0.485, 0.456, 0.406]).reshape(3, 1, 1) |
| 111 | self.std = torch.tensor([0.229, 0.224, 0.225]).reshape(3, 1, 1) |
| 112 | self.length = info[dataset][split] |
| 113 | self.env = None |
| 114 | def _init_db(self): |
| 115 | self.env = lmdb.open(self.lmdb_dir, |
| 116 | subdir=os.path.isdir(self.lmdb_dir), |
| 117 | readonly=True, |
| 118 | lock=False, |
| 119 | readahead=False, |
| 120 | meminit=False) |
| 121 | with self.env.begin(write=False) as txn: |
| 122 | self.length = loads_pyarrow(txn.get(b'__len__')) |
| 123 | self.keys = loads_pyarrow(txn.get(b'__keys__')) |
| 124 | |
| 125 | def __len__(self): |
| 126 | return self.length |
| 127 | |
| 128 | def __getitem__(self, index): |
| 129 | # Delay loading LMDB data until after initialization: https://github.com/chainer/chainermn/issues/129 |
| 130 | if self.env is None: |
| 131 | self._init_db() |
| 132 | env = self.env |
| 133 | with env.begin(write=False) as txn: |
| 134 | byteflow = txn.get(self.keys[index]) |
| 135 | ref = loads_pyarrow(byteflow) |
| 136 | # img |
| 137 | ori_img = cv2.imdecode(np.frombuffer(ref['img'], np.uint8), |
| 138 | cv2.IMREAD_COLOR) |
| 139 | img = cv2.cvtColor(ori_img, cv2.COLOR_BGR2RGB) |
| 140 | img_size = img.shape[:2] |
| 141 | # mask |
| 142 | seg_id = ref['seg_id'] |
| 143 | mask_dir = os.path.join(self.mask_dir, str(seg_id) + '.png') |
| 144 | # sentences |
| 145 | idx = np.random.choice(ref['num_sents']) |
| 146 | sents = ref['sents'] |
| 147 | # transform |
| 148 | # mask transform |
| 149 | mask = cv2.imdecode(np.frombuffer(ref['mask'], np.uint8), |
| 150 | cv2.IMREAD_GRAYSCALE) |
| 151 | mask = mask / 255. |
| 152 | if self.mode == 'train': |
| 153 | sent = sents[idx] |
| 154 | # sentence -> vector |
| 155 | img, mask, sent = self.convert(img, mask, sent, inference=False) |
nothing calls this directly
no outgoing calls
no test coverage detected