File size: 1,096 Bytes
324d5f8 cbd5880 324d5f8 cbd5880 324d5f8 cbd5880 324d5f8 cbd5880 324d5f8 60ff7a4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 |
---
library_name: xpmir
---
# monoBERT trained on MS-Marco
Passage Re-ranking with BERT (Rodrigo Nogueira, Kyunghyun Cho). 2019.
https://arxiv.org/abs/1901.04085
This model has been trained on MsMarco v1
## Using the model
The model can be loaded with [experimaestro
IR](https://experimaestro-ir.readthedocs.io/en/latest/)
```py from xpmir.models import AutoModel
from xpmir.models import AutoModel
# Model that can be re-used in experiments
model = AutoModel.load_from_hf_hub("xpmir/monobert")
# Use this if you want to actually use the model
model = AutoModel.load_from_hf_hub("xpmir/monobert", as_instance=True)
model.initialize()
model.rsv("walgreens store sales average", "The average Walgreens salary ranges...")
```
## Results
| Dataset | AP | P@20 | RR | RR@10 | nDCG | nDCG@10 | nDCG@20 |
|----| ---|------|------|------|------|------|------|
| msmarco_dev | 0.3563 | 0.0367 | 0.3611 | 0.3515 | 0.4626 | 0.4127 | 0.4344 |
| trec2019 | 0.4971 | 0.7163 | 0.9535 | 0.9535 | 0.6909 | 0.7081 | 0.6820 |
| trec2020 | 0.4763 | 0.6120 | 0.9093 | 0.9080 | 0.6797 | 0.6816 | 0.6540 |
|