Joctor/newyorker_oogiri_caption
收藏Hugging Face2024-06-13 更新2024-06-12 收录
下载链接：
https://hf-mirror.com/datasets/Joctor/newyorker_oogiri_caption
下载链接
链接失效反馈官方服务：
资源简介：
---
dataset_info:
- config_name: default
features:
- name: image
dtype: image
- name: contest_number
dtype: int64
- name: captions
sequence: string
- name: mean
sequence: float64
splits:
- name: train
num_bytes: 359384521.0
num_examples: 365
download_size: 294881876
dataset_size: 359384521.0
- config_name: matching
features:
- name: image
dtype: image
- name: contest_number
dtype: int64
- name: caption_choices
sequence: string
- name: label
dtype: string
splits:
- name: train
num_bytes: 1334473798.0
num_examples: 29200
- name: valid
num_bytes: 134234764.0
num_examples: 3600
- name: test
num_bytes: 819618978.0
num_examples: 3700
download_size: 284043330
dataset_size: 2288327540.0
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- config_name: matching
data_files:
- split: train
path: matching/train-*
- split: valid
path: matching/valid-*
- split: test
path: matching/test-*
---
The dataset includes two configurations: default and matching. The default configuration features images, contest numbers, captions, and means, with a training set of 365 samples. The matching configuration includes images, contest numbers, caption choices, and labels, with training, validation, and test sets containing 29200, 3600, and 3700 samples respectively.
提供机构：
Joctor
