README.md

metadata

license: mit
dataset_info:
  features:
    - name: _id
      dtype: string
    - name: title
      dtype: string
    - name: text
      dtype: string
    - name: openai
      sequence: float32
  splits:
    - name: train
      num_bytes: 12383152
      num_examples: 1000000
  download_size: 12383152
  dataset_size: 1000000
language:
  - en
pretty_name: OpenAI 1M with DBPedia Entities
size_categories:
  - 1M<n<10M

1M OpenAI Embeddings (1536 dimensions) from June 2023.

Text used for Embedding: title (string) + text (string)

First used for the pgvector vs VectorDB (Qdrant) benchmark: https://nirantk.com/writing/pgvector-vs-qdrant/

Future work

We are planning to take this up to 10M (and possibly 100M) vectors. Contact @KShivendu_ if you want to help :)

Credits:

This dataset was generated from the first 1M entries of https://huggingface.co/datasets/BeIR/dbpedia-entity