Update README.md

README.md

@@ -4,15 +4,42 @@ library_name: keras
 
 ## Model description
 
-More information needed
+BERT-based model for predicting fake news written in Romanian.
 
 ## Intended uses & limitations
 
-More information needed
+It predicts one of six types of fake news (in order: "fabricated", "fictional", "plausible", "propaganda", "real", "satire").
 
-## Training and evaluation data
+It also predicts if the article talks about health or politcs.
 
-More information needed
+## How to use
 
- ## Training Metrics
-Model history needed
+Load the model with:
+
+```python
+from huggingface_hub import from_pretrained_keras
+
+model = from_pretrained_keras("pandrei7/fakenews-mtl")
+```
+
+Use this tokenizer: `readerbench/RoBERT-base`.
+
+The input length should be 512. You can tokenize the input like this:
+
+```python
+tokenizer(
+    your_text,
+    padding="max_length",
+    truncation=True,
+    max_length=512,
+    return_tensors="tf",
+)
+```
+
+## Training data
+
+The model was trained and evaluated on the [fakerom](https://www.tagtog.net/fakerom/fakerom) dataset.
+
+## Evaluation results
+
+The accuracy of predicting fake news was roughly 75%.