You can deploy this model and start earning money today!

rkmakd2258-wq/
bert-based-uncased-agnews4-v01

Name: bert-based-uncased-agnews4-v01
Availability: InStock

safetensors

bert

Views

No views yet

language:

en tags:
text-classification
ag-news
fine-tuning license: "mit" datasets:
ag_news metrics:
accuracy
f1 base_model: "bert-base-uncased"

Model Card for AG News 4-class Classifier

Model Overview

This model is a fine-tuned version of a pre-trained BERT model, adapted for 4-category news classification using the AG News dataset.
It predicts one of four news topics based on input text (news headline or article).

Task: Text classification
Input: English news text
Output: One of four category labels

Label	Category
0	World
1	Sports
2	Business
3	Science/Technology

Details

Base model: bert-base-uncased
Dataset: AG News
Framework: Hugging Face Transformers
Training epochs: 3
Batch size: 8
Optimizer: AdamW
Max length: 128

Usage

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification

model_name = "rkmakd2258-wq/agnews-bert-4class"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

text = "Apple announced its latest earnings report."
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)
pred = outputs.logits.argmax(dim=1).item()

labels = ["World", "Sports", "Business", "Science/Technology"]
print("Prediction:", labels[pred])

Metrics

(Update with actual results)

Metric	Value
Accuracy	-
F1 Score	-

Applications

News article categorization
Data preprocessing for news analytics
Media content filtering

Limitations

Trained mainly on English news articles from AG News.
May not generalize to other types of text without further fine-tuning.
Category mapping is fixed: 0=World, 1=Sports, 2=Business, 3=Science/Technology.

License

MIT

Author

Creator: rkmakd2258-wq
Contact: rkmakd2258@gmail.com

rkmakd2258-wq/bert-based-uncased-agnews4-v01