---
dataset_info:
- config_name: corpus
  features:
  - name: id
    dtype: string
  - name: modality
    dtype: string
  - name: text
    dtype: string
  - name: image
    dtype: image
  - name: vision
    dtype: image
  splits:
  - name: test
    num_bytes: 7627022
    num_examples: 800
  download_size: 4200044
  dataset_size: 7627022
- config_name: qrels
  features:
  - name: query-id
    dtype: string
  - name: corpus-id
    dtype: string
  - name: score
    dtype: int64
  splits:
  - name: test
    num_bytes: 5932
    num_examples: 200
  download_size: 4976
  dataset_size: 5932
- config_name: query
  features:
  - name: id
    dtype: string
  - name: modality
    dtype: string
  - name: text
    dtype: string
  - name: image
    dtype: image
  splits:
  - name: test
    num_bytes: 31443695
    num_examples: 200
  download_size: 31403760
  dataset_size: 31443695
configs:
- config_name: corpus
  data_files:
  - split: test
    path: corpus/test-*
- config_name: qrels
  data_files:
  - split: test
    path: qrels/test-*
- config_name: query
  data_files:
  - split: test
    path: query/test-*
license: mit
---

The evaluation code is implemented based on [MTEB](https://github.com/embeddings-benchmark/mteb/tree/main/mteb) framework and avaliable in https://github.com/rebeccaz4/MRMR.