This model is a multi-label transformer encoder based on dccuchile/bert-base-spanish-wwm-cased (BETO) fine-tuned on the official AEMPS CIMA Research Dataset (hsilvosa/aemps-cima) and integrated with procurement metadata from (hsilvosa/openplacsp).

Given a drug description, active ingredients, dosage form, or patient leaflet snippet in Spanish, it predicts its 5-level ATC (Anatomical Therapeutic Chemical) code taxonomy classification.

Intended Use & Disclaimer

Intended exclusively for medical research and pharmaceutical data analytics. Not intended for clinical prescribing or medical advice.

Usage

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("your-hf-username/BETO-ATC-Hierarchical-Classifier")
model = AutoModelForSequenceClassification.from_pretrained("your-hf-username/BETO-ATC-Hierarchical-Classifier")

text = "Medicamento: Omeprazol 20 mg | Forma farmacéutica: COMPRIMIDO | Principios activos: Omeprazol"
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)
probs = torch.softmax(outputs.logits, dim=-1)
print(probs)

hsilvosa/BETO-ATC-Hierarchical-Classifier

BETO ATC Hierarchical Classifier

Intended Use & Disclaimer

Usage

hsilvosa/
BETO-ATC-Hierarchical-Classifier