Domino-ai/amazon_polarity_10_pct

---
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
dataset_info:
features:
- name: label
dtype:
class_label:
names:
'0': negative
'1': positive
- name: title
dtype: string
- name: content
dtype: string
splits:
- name: train
num_bytes: 163359702
num_examples: 360000
- name: test
num_bytes: 18182813
num_examples: 40000
download_size: 120691417
dataset_size: 181542515
---

# Amazon Polarity 10pct

This is a direct subset of the original [Amazon Polarity](https://huggingface.co/datasets/amazon_polarity) dataset, downsampled 10pct with a random shuffle

### Dataset Summary

For quicker testing on Amazon Polarity. See https://huggingface.co/datasets/amazon_polarity for details and attributions

### Source Data
```python
from datasets import ClassLabel, Dataset, DatasetDict, load_dataset

ds_full = load_dataset("amazon_polarity", streaming=True)
ds_train_10_pct = Dataset.from_list(list(ds_full["train"].shuffle(seed=42).take(360_000)))
ds_test_10_pct = Dataset.from_list(list(ds_full["test"].shuffle(seed=42).take(40_000)))

ds_10_pct = DatasetDict({"train": ds_train_10_pct, "test": ds_test_10_pct})
# Need to recreate the class labels
class_label = ClassLabel(num_classes=2, names=["negative", "positive"])
ds_10_pct = ds_10_pct.map(lambda row: {"title": row["title"], "content": row["content"], "label": "negative" if not row["label"] else "positive"})
ds_10_pct = ds_10_pct.cast_column("label", class_label)
```

原始信息汇总

Amazon Polarity 10pct

数据集概述

该数据集是原始Amazon Polarity数据集的直接子集，随机抽样10%。

数据集配置

• 配置名称: default
• 数据文件:
  • 训练集: data/train-*
  • 测试集: data/test-*

数据集信息

• 特征:
  • label: 标签，数据类型为类别标签，包含两个类别：negative 和 positive。
  • title: 标题，数据类型为字符串。
  • content: 内容，数据类型为字符串。
• 分割:
  • 训练集:
    • 字节数: 163359702
    • 样本数: 360000
  • 测试集:
    • 字节数: 18182813
    • 样本数: 40000
• 下载大小: 120691417
• 数据集大小: 181542515