Modelpublic

chiunhau/news-classifier-1

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

0likes8downloads

Model Card

News Topic Classifier (BCE loss)

Fine-tuned `bert-base-uncased` for multi-label news topic classification using Binary Cross-Entropy loss.

Model details


Base model	`bert-base-uncased`
Task	Multi-label classification
Loss	Binary Cross-Entropy (BCE)
Number of labels	126 topic codes
Max input length	256 tokens (title + body)
Best val micro-F1	~0.890

Usage

python

from transformers import BertForSequenceClassification, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("chiunhau/news-classifier-1")
model     = BertForSequenceClassification.from_pretrained("chiunhau/news-classifier-1")
model.eval()

title = "Finland wins gold in ice hockey."
text  = "The Finnish national team claimed victory in the final."
inputs = tokenizer(title, text, return_tensors="pt",
                   truncation=True, max_length=256)
with torch.no_grad():
    logits = model(**inputs).logits

threshold = 0.5
probs     = logits.sigmoid().squeeze()
predicted = [model.config.id2label[i]
             for i, p in enumerate(probs) if p > threshold]
print("Topics:", predicted)

Notes

—Input: concatenated title + text fields, separated by the tokenizer's [SEP] token.
—Output: raw logits — apply sigmoid() and threshold at 0.5 for binary predictions.
—The companion ASL model (chiunhau/news-classifier-2) uses Asymmetric Loss and may be better at detecting rare topics.