Module `vipy.data.places`

Expand source code Browse git

import os
import vipy


URL = ['http://data.csail.mit.edu/places/places365/train_256_places365standard.tar',
       'http://data.csail.mit.edu/places/places365/val_256.tar']
MD5 = [None, None]


class Places365():
    """Project: http://places2.csail.mit.edu/download-private.html"""
    def __init__(self, datadir=None, redownload=False):

        datadir = vipy.util.tocache('places365') if datadir is None else datadir
        
        self._datadir = vipy.util.remkdir(datadir)
        for (url, md5) in zip(URL, MD5):
            if redownload or not os.path.exists(os.path.join(self._datadir, '.complete')):
                vipy.downloader.download_and_unpack(url, self._datadir, md5=md5)

        open(os.path.join(self._datadir, '.complete'), 'a').close()
        
    def trainset(self):
        imlist = tuple((f, vipy.util.filebase(vipy.util.filepath(f))) for f in vipy.util.findimages(os.path.join(self._datadir, 'data_256')))
        loader = lambda x: vipy.image.ImageCategory(filename=x[0], category=x[1])
        return vipy.dataset.Dataset(imlist, id='places365:train', loader=loader)

    def valset(self):
        imlist = tuple((f, vipy.util.filebase(vipy.util.filepath(f))) for f in vipy.util.findimages(os.path.join(self._datadir, 'val_256')))
        loader = lambda x: vipy.image.ImageCategory(filename=x[0], category=x[1])        
        return vipy.dataset.Dataset(imlist, id='places365:val', loader=loader)

Classes

class Places365 (datadir=None, redownload=False)

Project: http://places2.csail.mit.edu/download-private.html

Expand source code Browse git

class Places365():
    """Project: http://places2.csail.mit.edu/download-private.html"""
    def __init__(self, datadir=None, redownload=False):

        datadir = vipy.util.tocache('places365') if datadir is None else datadir
        
        self._datadir = vipy.util.remkdir(datadir)
        for (url, md5) in zip(URL, MD5):
            if redownload or not os.path.exists(os.path.join(self._datadir, '.complete')):
                vipy.downloader.download_and_unpack(url, self._datadir, md5=md5)

        open(os.path.join(self._datadir, '.complete'), 'a').close()
        
    def trainset(self):
        imlist = tuple((f, vipy.util.filebase(vipy.util.filepath(f))) for f in vipy.util.findimages(os.path.join(self._datadir, 'data_256')))
        loader = lambda x: vipy.image.ImageCategory(filename=x[0], category=x[1])
        return vipy.dataset.Dataset(imlist, id='places365:train', loader=loader)

    def valset(self):
        imlist = tuple((f, vipy.util.filebase(vipy.util.filepath(f))) for f in vipy.util.findimages(os.path.join(self._datadir, 'val_256')))
        loader = lambda x: vipy.image.ImageCategory(filename=x[0], category=x[1])        
        return vipy.dataset.Dataset(imlist, id='places365:val', loader=loader)

Methods

def trainset(self)

Expand source code Browse git

def trainset(self):
    imlist = tuple((f, vipy.util.filebase(vipy.util.filepath(f))) for f in vipy.util.findimages(os.path.join(self._datadir, 'data_256')))
    loader = lambda x: vipy.image.ImageCategory(filename=x[0], category=x[1])
    return vipy.dataset.Dataset(imlist, id='places365:train', loader=loader)

def valset(self)

Expand source code Browse git

def valset(self):
    imlist = tuple((f, vipy.util.filebase(vipy.util.filepath(f))) for f in vipy.util.findimages(os.path.join(self._datadir, 'val_256')))
    loader = lambda x: vipy.image.ImageCategory(filename=x[0], category=x[1])        
    return vipy.dataset.Dataset(imlist, id='places365:val', loader=loader)