Datasets:
metadata
dataset_info:
features:
- name: photo
struct:
- name: id
dtype: string
- name: url
dtype: string
- name: image_url
dtype: string
- name: submitted_at
dtype: string
- name: featured
dtype: bool
- name: width
dtype: uint16
- name: height
dtype: uint16
- name: aspect_ratio
dtype: float32
- name: description
dtype: string
- name: blur_hash
dtype: string
- name: photographer
struct:
- name: username
dtype: string
- name: first_name
dtype: string
- name: last_name
dtype: string
- name: exif
struct:
- name: camera_make
dtype: string
- name: camera_model
dtype: string
- name: iso
dtype: string
- name: aperture_value
dtype: string
- name: focal_length
dtype: string
- name: exposure_time
dtype: string
- name: location
struct:
- name: name
dtype: string
- name: latitude
dtype: float32
- name: longitude
dtype: float32
- name: country
dtype: string
- name: city
dtype: string
- name: stats
struct:
- name: views
dtype: uint32
- name: downloads
dtype: uint32
- name: ai
struct:
- name: description
dtype: string
- name: primary_landmark_name
dtype: string
- name: primary_landmark_latitude
dtype: string
- name: primary_landmark_longitude
dtype: string
- name: primary_landmark_confidence
dtype: string
- name: keywords
list:
- name: keyword
dtype: string
- name: ai_service_1_confidence
dtype: string
- name: ai_service_2_confidence
dtype: string
- name: suggested_by_user
dtype: bool
- name: collections
list:
- name: collection_id
dtype: string
- name: collection_title
dtype: string
- name: photo_collected_at
dtype: string
- name: conversions
list:
- name: converted_at
dtype: string
- name: conversion_type
dtype: string
- name: keyword
dtype: string
- name: anonymous_user_id
dtype: string
- name: conversion_country
dtype: string
- name: colors
list:
- name: hex
dtype: string
- name: red
dtype: uint8
- name: green
dtype: uint8
- name: blue
dtype: uint8
- name: keyword
dtype: string
- name: ai_coverage
dtype: float32
- name: ai_score
dtype: float32
splits:
- name: train
num_bytes: 1202216966
num_examples: 25000
download_size: 618337921
dataset_size: 1202216966
configs:
- config_name: default
data_files:
- split: train
path: data/train-*