---
language: en
license: mit
---
This is the baseline model used in most experiments in the paper ["A Dataset for N-ary Relation Extraction of Drug Combinations"](https://arxiv.org/abs/2205.02289).

*(for just the domain-adapted masked language model that we use underneath this model, see [here](https://huggingface.co/allenai/drug_combinations_lm_pubmedbert?text=Paxlovid+works+well+in+combination+with+%5BMASK%5D+for+treating+breast+cancer.))*


**Steps to load this model**

1) Download accompanying code:
```
git clone https://github.com/allenai/drug-combo-extraction.git
conda create --name drug_combo python=3.8.5
conda activate drug_combo
```

2) Download model from Huggingface:

```
git lfs install
git clone https://huggingface.co/allenai/drug-combo-classifier-pubmedbert-dapt
```

3) Load model (`in Python`):

```
from modeling.model import load_model
checkpoint_path = "drug-combo-classifier-pubmedbert-dapt"
model, tokenizer, metadata = load_model(checkpoint_path)
```