Update README.md

# BERT Model for Romanian Fake News Detection

## Model Information

- **Model Name**: bert_model_enhanced_2
- **Model Path**: ./bert_model_enhanced_2
- **Architecture**: BertForSequenceClassification
- **Model Type**: bert

## Hyperparameters

- **Attention Probabilities Dropout Probability**: 0.1
- **Classifier Dropout**: null
- **Directionality**: bidi
- **Hidden Activation Function**: gelu
- **Hidden Dropout Probability**: 0.1
- **Hidden Size**: 768
- **Intermediate Size**: 3072
- **Initializer Range**: 0.02
- **Layer Normalization Epsilon**: 1e-12
- **Maximum Position Embeddings**: 512
- **Number of Attention Heads**: 12
- **Number of Hidden Layers**: 12

## Label Mappings

- **ID to Label**:
- 0: fake_news
- 1: misinformations
- 2: propaganda
- 3: real_news
- 4: satire


## Usage

This model is trained to detect Romanian fake news. It can be used for sequence classification tasks where the input is a piece of text in Romanian, and the output is one of the predefined labels indicating the likelihood of the text being fake news.

README.md

@@ -1,13 +1,14 @@
----
-license: mit
-language:
-- ro
-metrics:
-- accuracy
-base_model: google-bert/bert-base-uncased
-library_name: transformers.js
-tags:
-- romanian
-- fake news
-- BERT
----
+---
+license: mit
+language:
+- ro
+metrics:
+- accuracy
+base_model: google-bert/bert-base-uncased
+library_name: transformers.js
+tags:
+- romanian
+- fake news
+- BERT
+---
+