langformers/allnli-mimic-embedding

---
task_categories:
- feature-extraction
language:
- en
size_categories:
- 100K<n<1M
---
# ALLNLI for Mimicking Vector Space

## Description
This dataset contains sentences from the [ALLNLI](https://huggingface.co/datasets/sentence-transformers/all-nli) dataset, which is a combination of the [SNLI](https://huggingface.co/datasets/stanfordnlp/snli) and [MultiNLI](https://huggingface.co/datasets/nyu-mll/multi_nli) datasets. It is designed for training a student model to mimic the vector space of a teacher model. This dataset is particularly useful for tasks involving embedding loss computation, where the student model learns to replicate the teacher model's embeddings. All the "anchor" and "positive" sentences from the ALLNLI dataset are included, with redundant sentences removed.

## Usage
You can load the dataset using the Hugging Face `datasets` library:

```python
from datasets import load_dataset

dataset = load_dataset('langformers/allnli-mimic-embedding')

---
任务类别：
- 特征提取（feature-extraction）
语言：
- 英语
数据规模：
- 10万 < 样本量 < 100万
---

# 用于向量空间模仿的ALLNLI数据集

## 数据集说明
本数据集的语句取自[ALLNLI](https://huggingface.co/datasets/sentence-transformers/all-nli)数据集，该数据集是[SNLI](https://huggingface.co/datasets/stanfordnlp/snli)与[MultiNLI](https://huggingface.co/datasets/nyu-mll/multi_nli)数据集的整合集合。本数据集旨在训练学生模型以模仿教师模型的向量空间，尤其适用于涉及嵌入损失（embedding loss）计算的任务：学生模型可通过该数据集学习复现教师模型的嵌入向量。本数据集保留了ALLNLI数据集中所有的「锚点（anchor）」与「正例（positive）」语句，并已移除冗余语句。

## 使用方法
用户可通过Hugging Face的`datasets`库加载本数据集：
python
from datasets import load_dataset

dataset = load_dataset('langformers/allnli-mimic-embedding')