Update README.md

README.md

@@ -15,49 +15,95 @@ datasets:
 - custom
 ---
 
-# PIDIT: Modello Multi-Task BERT + ALBERTO per analisi ideologica e di genere 🇮🇹
+# PIDIT: Multi-Task BERT + ALBERTO Model for Gender and Ideology Prediction 🇮🇹
 
-Questo modello `tf.keras` unisce due encoder pre-addestrati (`BERT` e `ALBERTO`) per effettuare predizioni multi-task su testi in italiano.  
-È progettato per classificare:
+This `tf.keras` model combines two pre-trained encoders — `BERT` and `ALBERTO` — to perform multi-task classification on Italian-language texts.  
+It is designed to predict:
 
-- 🧑‍🤝‍🧑 **Genere** dell'autore (binary classification)
-- 🏛️ **Ideologia binaria** (es. conservatore vs progressista)
-- 🧭 **Ideologia multiclasse** (4 classi ideologiche)
+- 🧑‍🤝‍🧑 **Author gender** (binary classification)
+- 🏛️ **Binary ideology** (e.g., progressive vs conservative)
+- 🧭 **Multiclass ideology** (4 ideological classes)
 
-## ✨ Architettura
+## ✨ Architecture
 
-- `TFBertModel` da `bert-base-italian-uncased` (non fine-tuned)
-- `TFAutoModel` da `alberto-base-uncased` (non fine-tuned)
-- Layer di concatenazione e densi condivisi
-- 3 teste di output:
+- `TFBertModel` from `bert-base-italian-uncased` (frozen)
+- `TFAutoModel` from `alberto-base-uncased` (frozen)
+- Concatenated outputs + dense layers
+- Three output heads:
   - `gender`: `Dense(1, activation="sigmoid")`
   - `ideology_binary`: `Dense(1, activation="sigmoid")`
   - `ideology_multiclass`: `Dense(4, activation="softmax")`
 
 ## 📥 Input
 
-Il modello accetta **6 input**:
+The model takes **6 input tensors**:
 - `bert_input_ids`, `bert_token_type_ids`, `bert_attention_mask`
 - `alberto_input_ids`, `alberto_token_type_ids`, `alberto_attention_mask`
 
-Tutti con shape `(batch_size, max_length)`.
+All tensors have shape `(batch_size, max_length)`.
 
 ---
 
-## 🚀 Utilizzo
+## 🚀 Usage
 
-### 1. Caricamento del modello
+### 1. Load the model
 
 ```python
 from huggingface_hub import snapshot_download
 from transformers import TFBertModel, TFAutoModel
 import tensorflow as tf
 
-# Scarica localmente il modello
+# Download the model locally
 model_path = snapshot_download("leeeov4/PIDIT")
 
-# Carica il modello
+# Load the model
 model = tf.keras.models.load_model(model_path, custom_objects={
     "TFBertModel": TFBertModel,
     "TFAutoModel": TFAutoModel
 })
+```
+
+### 2. Load the tokenizers
+
+```python
+from transformers import AutoTokenizer
+
+bert_tokenizer = AutoTokenizer.from_pretrained("leeeov4/PIDIT/bert_tokenizer")
+alberto_tokenizer = AutoTokenizer.from_pretrained("leeeov4/PIDIT/alberto_tokenizer")
+```
+
+## 🧼 Preprocessing Example
+
+```python
+def preprocess_text(text, max_length=250):
+    bert_tokens = bert_tokenizer(text, max_length=max_length, padding='max_length', truncation=True, return_tensors='tf')
+    alberto_tokens = alberto_tokenizer(text, max_length=max_length, padding='max_length', truncation=True, return_tensors='tf')
+
+    return {
+        'bert_input_ids': bert_tokens['input_ids'],
+        'bert_token_type_ids': bert_tokens['token_type_ids'],
+        'bert_attention_mask': bert_tokens['attention_mask'],
+        'alberto_input_ids': alberto_tokens['input_ids'],
+        'alberto_token_type_ids': alberto_tokens['token_type_ids'],
+        'alberto_attention_mask': alberto_tokens['attention_mask']
+    }
+
+```
+
+
+## 🧼 Inference
+
+```python
+text = "Questo è un esempio di testo italiano per testare il modello."
+inputs = preprocess_text(text)
+outputs = model.predict(inputs)
+
+gender_prob = outputs[0][0][0]
+ideology_binary_prob = outputs[1][0][0]
+ideology_multiclass_probs = outputs[2][0]
+
+print("Predicted gender (male probability):", gender_prob)
+print("Predicted binary ideology (conservative probability):", ideology_binary_prob)
+print("Multiclass ideology distribution:", ideology_multiclass_probs)
+
+```