Example parsing inaturalist dataset

parse_inat_dataset_ex.py

import cPickle as pickle
import os

dataset_dir = '.'
train_file = os.path.join(dataset_dir, "train_data.pkl")
with open(train_file) as f:
  train_data = pickle.load(f)

image_dir = "/home/gvanhorn/datasets/inaturalist/images"
train_images = []
train_labels = []
for image in train_data:
  fn = image['filename']
  label = image['label']
  
  train_images.append(os.path.join(image_dir, fn))
  train_labels.append(label)

# Now do something...