Created
August 6, 2018 08:30
-
-
Save chenyaofo/d6148b2ff6e85fdf40f983109f64326d to your computer and use it in GitHub Desktop.
MIODataset
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import copy | |
from torch.utils.data import Dataset | |
from torchlearning.mio import MIO, Split | |
class MioDataset(Dataset): | |
def __init__(self, root, sampler, transform=None, target_transform=None): | |
self.root = root | |
self.sampler = sampler | |
self.transform = transform | |
self.target_transform = target_transform | |
self.mio = MIO(self.root) | |
self.split = None | |
def to_split(self, split: Split): | |
dataset = copy.copy(self) | |
dataset.split = split.items | |
return dataset | |
def __getitem__(self, id_): | |
if self.split is not None: | |
id_ = self.split[id_] | |
size = self.mio.get_collection_size(id_) | |
selected_samples = self.sampler(size) | |
if isinstance(selected_samples, int): | |
object_id = selected_samples | |
data = self.mio.fetchone(id_, object_id) | |
else: | |
data = self.mio.fetchmany(id_, selected_samples) | |
target = self.mio.get_collection_metadata(id_) | |
if self.transform is not None: | |
data = self.transform(data) | |
if self.target_transform is not None: | |
target = self.target_transform(target) | |
return data, target | |
def __len__(self): | |
if self.split is None: | |
return self.mio.size | |
else: | |
return len(self.split) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment