---
license: mit
pipeline_tag: text-ranking
tags:
- mteb
---

# bm25s


Repo for storing evaluation results of the [`bm25s`](https://embeddings-benchmark.github.io/mteb/overview/available_models/text/?h=bm25#baselinebm25s)
implementation in MTEB. BM25 acts as a competitive hard-to-beat baseline.

Note that this repository does not contain the model implementation, but you can fetch the implementation in MTEB using:

```python
import mteb

task = mteb.get_benchmark("RTEB(beta)")  # the v2 English benchmark
model = mteb.get_model("mteb/baseline-bm25s")

results = mteb.evaluate(model, task)
```