| | --- |
| | 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') |
