Commit
•
383ea9a
1
Parent(s):
b3db12e
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
## Model Details
|
2 |
+
This model is a Binart classification model fine-tuned on the FakeOrRealNews Dataset using the BERT (bert-base-uncased) architecture. The primary task is to classify news articles into different categories, making it suitable for fake news detection. BERT (Bidirectional Encoder Representations from Transformers) is a transformer-based model known for its effectiveness in natural language processing tasks.
|
3 |
+
|
4 |
+
It takes the title of the news article and classifies it into Reliable or Unreliable news.
|
5 |
+
|
6 |
+
Bias: The model may inherit biases present in the training data, and it's important to be aware of potential biases in the predictions.
|
7 |
+
|
8 |
+
## Code Implementation
|
9 |
+
```python
|
10 |
+
# Load model directly
|
11 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
12 |
+
|
13 |
+
tokenizer = AutoTokenizer.from_pretrained("Arjun24420/BERT-FakeOrReal-BinaryClassification")
|
14 |
+
model = AutoModelForSequenceClassification.from_pretrained("Arjun24420/BERT-FakeOrReal-BinaryClassification")
|
15 |
+
|
16 |
+
def predict(text):
|
17 |
+
# Tokenize the input text and move tensors to the GPU if available
|
18 |
+
inputs = tokenizer(text, padding=True, truncation=True,
|
19 |
+
max_length=512, return_tensors="pt")
|
20 |
+
|
21 |
+
# Get model output (logits)
|
22 |
+
outputs = model(**inputs)
|
23 |
+
|
24 |
+
probs = outputs.logits.softmax(1)
|
25 |
+
# Get the probabilities for each class
|
26 |
+
class_probabilities = {class_mapping[i]: probs[0, i].item()
|
27 |
+
for i in range(probs.shape[1])}
|
28 |
+
|
29 |
+
return class_probabilities
|
30 |
+
|
31 |
+
```
|