---
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: 49188465
    num_examples: 700
  download_size: 47822991
  dataset_size: 49188465
- config_name: qrels
  features:
  - name: query-id
    dtype: string
  - name: corpus-id
    dtype: string
  - name: score
    dtype: int64
  splits:
  - name: test
    num_bytes: 2067
    num_examples: 88
  download_size: 2316
  dataset_size: 2067
- 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: 84333786
    num_examples: 88
  download_size: 57852600
  dataset_size: 84333786
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.