jxu124/refcocoplus

---
dataset_info:
features:
- name: sent_ids
sequence: int64
- name: file_name
dtype: string
- name: ann_id
dtype: int64
- name: ref_id
dtype: int64
- name: image_id
dtype: int64
- name: split
dtype: string
- name: sentences
list:
- name: raw
dtype: string
- name: sent
dtype: string
- name: sent_id
dtype: int64
- name: tokens
sequence: string
- name: category_id
dtype: int64
- name: raw_anns
dtype: string
- name: raw_image_info
dtype: string
- name: raw_sentences
dtype: string
- name: image_path
dtype: string
- name: bbox
sequence: float64
- name: captions
sequence: string
- name: global_image_id
dtype: string
- name: anns_id
dtype: string
splits:
- name: train
num_bytes: 81937869
num_examples: 42278
- name: testB
num_bytes: 3273927
num_examples: 1798
- name: test
num_bytes: 3969265
num_examples: 1975
- name: validation
num_bytes: 7399541
num_examples: 3805
download_size: 39772801
dataset_size: 96580602
---
# Dataset Card for "refcocoplus"

[More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)

原始信息汇总

数据集概述

特征信息

• sent_ids: 序列类型，数据类型为 int64
• file_name: 数据类型为 string
• ann_id: 数据类型为 int64
• ref_id: 数据类型为 int64
• image_id: 数据类型为 int64
• split: 数据类型为 string
• sentences: 列表类型，包含以下子特征：
  • raw: 数据类型为 string
  • sent: 数据类型为 string
  • sent_id: 数据类型为 int64
  • tokens: 序列类型，数据类型为 string
• category_id: 数据类型为 int64
• raw_anns: 数据类型为 string
• raw_image_info: 数据类型为 string
• raw_sentences: 数据类型为 string
• image_path: 数据类型为 string
• bbox: 序列类型，数据类型为 float64
• captions: 序列类型，数据类型为 string
• global_image_id: 数据类型为 string
• anns_id: 数据类型为 string

数据分割

• train: 字节数为 81937869，样本数为 42278
• testB: 字节数为 3273927，样本数为 1798
• test: 字节数为 3969265，样本数为 1975
• validation: 字节数为 7399541，样本数为 3805

数据集大小

• 下载大小: 39772801 字节
• 数据集大小: 96580602 字节