metadata
license: mit
dataset_info:
features:
- name: image
dtype: image
- name: Latitude
dtype: float64
- name: Longitude
dtype: float64
splits:
- name: train
num_bytes: 768821042
num_examples: 216
download_size: 768821042
dataset_size: 768821042
configs:
- config_name: default
data_files:
- split: train
path: data/train_*
Dataset Card for Dataset Name
Train_1 and train_3 contain images extracted from videos, while train_2, train_4, train_5, and train_6 contain images that were captured directly.
How each sample looks like:
{ "image": { "bytes": , "path": None }, "Latitude": 39.9524, "Longitude": -75.1912 }
Uses
Used to train our model for IMG2GPS