PyTorchClassification/data_loader.py [441:568]: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - else: self.targets = self.classes # print out some stats print ('\t' + str(len(self.imgs)) + ' images') print ('\t' + str(len(set(self.classes))) + ' classes') self.root = root self.is_train = is_train self.loader = ImageLoader(image_sizes) self.multi_crop = multi_crop self.imagesInZip = (".zip" in self.root) if (self.imagesInZip): self.archive = zipfile.ZipFile(self.root, 'r') numToUse = int((len(self.imgs)*percentUse)/100) self.imgs = self.imgs[:numToUse] self.ids = self.ids[:numToUse] self.classes = self.classes[:numToUse] if bbox_predictions: img_id_to_idx = {image_id:idx for idx,image_id in enumerate(self.ids)} self.bboxes = [None for _ in range(len(self.ids))] self.bbox_labels = [None for _ in range(len(self.ids))] self.bbox_scores = [None for _ in range(len(self.ids))] loaded_dict = np.load(bbox_predictions) image_ids = loaded_dict['image_ids'] for image_id, bbox, bbox_labels, bbox_scores in zip(image_ids, loaded_dict['pred_bboxes'], loaded_dict['pred_labels'], loaded_dict['pred_scores']): if len(bbox) > 0: assert image_id[0].tolist() in img_id_to_idx, 'Didn\'t find image for bounding box, ' + \ 'maybe it\'s the wrong json file?' idx = img_id_to_idx[image_id[0].tolist()] self.bboxes[idx] = bbox self.bbox_labels[idx] = bbox_labels self.bbox_scores[idx] = bbox_scores else: self.bboxes = None self.bbox_labels = None self.bbox_scores = None self.label_smoothing = label_smoothing def distanceMatrix(self): D = np.zeros((len(self.classes_taxonomic),len(self.classes_taxonomic))) norm = 1/len(self.tax_levels) for i in range(len(self.classes_taxonomic)): print("%d" % i, end=' ') for j in range(i, len(self.classes_taxonomic)): eq = np.equal(self.classes_taxonomic[i],self.classes_taxonomic[j]) D[i,j] = np.argmax(eq)*norm D[j,i] = D[i,j] return D def load_taxonomy(self, ann_data, tax_levels, classes): # loads the taxonomy data and converts to ints taxonomy = {} classnames = {} if 'categories' in ann_data.keys(): num_classes = len(ann_data['categories']) for tt in tax_levels: tax_data = [aa[tt] for aa in ann_data['categories']] _, tax_id = np.unique(tax_data, return_inverse=True) taxonomy[tt] = dict(zip(range(num_classes), list(tax_id))) for cat in ann_data['categories']: classnames[cat['id']] = cat['name'] else: # set up dummy data for tt in tax_levels: taxonomy[tt] = dict(zip([0], [0])) # create a dictionary of lists containing taxonomic labels classes_taxonomic = {} for cc in np.unique(classes): tax_ids = [0]*len(tax_levels) for ii, tt in enumerate(tax_levels): tax_ids[ii] = taxonomy[tt][cc] classes_taxonomic[tax_ids[0]] = tax_ids return taxonomy, classes_taxonomic, classnames def __getitem__(self, index): im_id = self.ids[index] species_id = self.targets[index] if self.bboxes is not None and self.bboxes[index] is not None: # bbox_scores = self.bbox_scores[index] # most_confident_bbox = np.argmax(bbox_scores) # bbox = self.bboxes[index][most_confident_bbox,:] bboxes = self.bboxes[index] else: bboxes = [] if self.imagesInZip: path = archive.open(self.imgs[index]) else: path = self.root + self.imgs[index] raw_image = self.loader.load_image(path) imgs = self.loader.process_image(raw_image, self.is_train, self.multi_crop, bboxes) return imgs, im_id, species_id def __len__(self): return len(self.imgs) def get_labels(self): return self.classes def get_num_classes(self): assert np.max(self.classes) < len(self.classnames) return len(self.classnames) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - PyTorchClassification/data_loader_cv.py [392:520]: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - else: self.targets = self.classes # print out some stats print ('\t' + str(len(self.imgs)) + ' images') print ('\t' + str(len(set(self.classes))) + ' classes') self.root = root self.is_train = is_train self.loader = ImageLoader(image_sizes) self.multi_crop = multi_crop self.imagesInZip = (".zip" in self.root) if (self.imagesInZip): self.archive = zipfile.ZipFile(self.root, 'r') numToUse = int((len(self.imgs)*percentUse)/100) self.imgs = self.imgs[:numToUse] self.ids = self.ids[:numToUse] self.classes = self.classes[:numToUse] if bbox_predictions: img_id_to_idx = {image_id:idx for idx,image_id in enumerate(self.ids)} self.bboxes = [None for _ in range(len(self.ids))] self.bbox_labels = [None for _ in range(len(self.ids))] self.bbox_scores = [None for _ in range(len(self.ids))] loaded_dict = np.load(bbox_predictions) image_ids = loaded_dict['image_ids'] for image_id, bbox, bbox_labels, bbox_scores in zip(image_ids, loaded_dict['pred_bboxes'], loaded_dict['pred_labels'], loaded_dict['pred_scores']): if len(bbox) > 0: assert image_id[0].tolist() in img_id_to_idx, 'Didn\'t find image for bounding box, ' + \ 'maybe it\'s the wrong json file?' idx = img_id_to_idx[image_id[0].tolist()] self.bboxes[idx] = bbox self.bbox_labels[idx] = bbox_labels self.bbox_scores[idx] = bbox_scores else: self.bboxes = None self.bbox_labels = None self.bbox_scores = None self.label_smoothing = label_smoothing def distanceMatrix(self): D = np.zeros((len(self.classes_taxonomic),len(self.classes_taxonomic))) norm = 1/len(self.tax_levels) for i in range(len(self.classes_taxonomic)): print("%d" % i, end=' ') for j in range(i, len(self.classes_taxonomic)): eq = np.equal(self.classes_taxonomic[i],self.classes_taxonomic[j]) D[i,j] = np.argmax(eq)*norm D[j,i] = D[i,j] return D def load_taxonomy(self, ann_data, tax_levels, classes): # loads the taxonomy data and converts to ints taxonomy = {} classnames = {} if 'categories' in ann_data.keys(): num_classes = len(ann_data['categories']) for tt in tax_levels: tax_data = [aa[tt] for aa in ann_data['categories']] _, tax_id = np.unique(tax_data, return_inverse=True) taxonomy[tt] = dict(zip(range(num_classes), list(tax_id))) for cat in ann_data['categories']: classnames[cat['id']] = cat['name'] else: # set up dummy data for tt in tax_levels: taxonomy[tt] = dict(zip([0], [0])) # create a dictionary of lists containing taxonomic labels classes_taxonomic = {} for cc in np.unique(classes): tax_ids = [0]*len(tax_levels) for ii, tt in enumerate(tax_levels): tax_ids[ii] = taxonomy[tt][cc] classes_taxonomic[tax_ids[0]] = tax_ids return taxonomy, classes_taxonomic, classnames def __getitem__(self, index): im_id = self.ids[index] species_id = self.targets[index] if self.bboxes is not None and self.bboxes[index] is not None: # bbox_scores = self.bbox_scores[index] # most_confident_bbox = np.argmax(bbox_scores) # bbox = self.bboxes[index][most_confident_bbox,:] bboxes = self.bboxes[index] else: bboxes = [] if self.imagesInZip: path = archive.open(self.imgs[index]) else: path = self.root + self.imgs[index] raw_image = self.loader.load_image(path) imgs = self.loader.process_image(raw_image, self.is_train, self.multi_crop, bboxes) return imgs, im_id, species_id def __len__(self): return len(self.imgs) def get_labels(self): return self.classes def get_num_classes(self): assert np.max(self.classes) < len(self.classnames) return len(self.classnames) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -