File size: 714 Bytes
edee94e
1b298f4
edee94e
 
 
 
1b298f4
edee94e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

---
license: mit
base_model: emilyalsentzer/Bio_ClinicalBERT
pipeline_tag: text-classification
tags: [multi-label, clinical, overdose]
---

# DrugDetectionBERT

Bio_ClinicalBERT fine-tuned for multi-label drug detection in death investigation narratives.

Labels: Methamphetamine, Heroin, Cocaine, Fentanyl, Alcohol, Prescription.opioids,
Any Opioids, Benzodiazepines, Others, Any Drugs.

Sigmoid outputs. Per-label thresholds in `best_thresholds.json` — do not use 0.5.

```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tok = AutoTokenizer.from_pretrained("Elemets/DrugDetectionBERT")
model = AutoModelForSequenceClassification.from_pretrained("Elemets/DrugDetectionBERT")