Update README.md

README.md

@@ -59,19 +59,19 @@ base_model:
 - Kelvinmbewe/mbert_LusakaLang_Topic
 ---
 
-## **LusakaLang Multi‑Task Model (Language + Sentiment + Topic)**
+## **LusakaLang MultiTask Model**
 
-This model is a unified transformer architecture built on top of **`bert-base-multilingual-cased`**, designed to perform **three tasks simultaneously**:
+This model is a unified transformer architecture built on top of `bert-base-multilingual-cased`, designed to perform three tasks simultaneously:
 
-1. **[Language Identification](guide://action?prefill=Tell%20me%20more%20about%3A%20Language%20Identification)**  
-2. **[Sentiment Analysis](guide://action?prefill=Tell%20me%20more%20about%3A%20Sentiment%20Analysis)**  
-3. **[Topic Classification](guide://action?prefill=Tell%20me%20more%20about%3A%20Topic%20Classification)**  
+1. Language Identification
+2. Sentiment Analysis
+3. Topic Classification
 
 The system integrates three fine‑tuned LusakaLang checkpoints:
 
-- **[Kelvinmbewe/mbert_Lusaka_Language_Analysis](guide://action?prefill=Tell%20me%20more%20about%3A%20Kelvinmbewe%2Fmbert_Lusaka_Language_Analysis)**  
-- **[Kelvinmbewe/mbert_LusakaLang_Sentiment_Analysis](guide://action?prefill=Tell%20me%20more%20about%3A%20Kelvinmbewe%2Fmbert_LusakaLang_Sentiment_Analysis)**  
-- **[Kelvinmbewe/mbert_LusakaLang_Topic](guide://action?prefill=Tell%20me%20more%20about%3A%20Kelvinmbewe%2Fmbert_LusakaLang_Topic)**  
+- mbert_Lusaka_Language_Analysis
+- mbert_LusakaLang_Sentiment_Analysis
+- mbert_LusakaLang_Topic
 
 All tasks share a single mBERT encoder, supported by three independent classifier heads. This architecture enhances computational efficiency, reduces memory overhead 
 and promotes consistent, harmonized predictions across all tasks.
@@ -92,20 +92,6 @@ understanding of real Zambian communication.
 
 ---
 
-# **Training Architecture**
-
-The model uses:
-
-- **Shared Encoder:** mBERT  
-- **Head 1:** Language classifier  
-- **Head 2:** Sentiment classifier  
-- **Head 3:** Topic classifier  
-
-This multi‑task setup improves generalization and reduces inference cost.
-
----
-
-
 ## **How to Use This Model**
 
 
@@ -114,46 +100,25 @@ from transformers import AutoTokenizer
 import torch
 
 class LusakaLangMultiTask:
-    def __init__(self, model_path="Kelvinmbewe/LusakaLang-MultiTask"):
-        self.tokenizer = AutoTokenizer.from_pretrained(model_path)
-        self.model = torch.load(f"{model_path}/model.pt")
-        self.model.eval()
-    def predict_language(self, texts):
-        # Your actual implementation goes here
-        pass
-    def predict_sentiment(self, texts):
-        # Your actual implementation goes here
-        pass
-    def predict_topic(self, texts):
-        # Your actual implementation goes here
-        pass
-
-# Instantiate model
+    def __init__(self, path="Kelvinmbewe/LusakaLang-MultiTask"):
+        self.tokenizer = AutoTokenizer.from_pretrained(path)
+        self.model = torch.load(f"{path}/model.pt").eval()
+
+    def predict_language(self, texts): pass
+    def predict_sentiment(self, texts): pass
+    def predict_topic(self, texts): pass
+
 llm = LusakaLangMultiTask()
-# Run predictions
-language_results = llm.predict_language([
-    "Ndeumfwa bwino lelo",
-    "Galimoto inachedwa koma driver anali bwino",
-    "The service was terrible today"
-])
-sentiment_results = llm.predict_sentiment([
-    "Driver was rude and unprofessional",
-    "Ndimvela bwino lelo",
-    "The ride was okay, nothing special"
-])
-topic_results = llm.predict_topic([
-    "Payment failed but money was deducted",
-    "Support siyankhapo, waited long",
-    "Driver was over speeding"
-])
-print(language_results)
-print(sentiment_results)
-print(topic_results)
+
+print(llm.predict_language([...]))
+print(llm.predict_sentiment([...]))
+print(llm.predict_topic([...]))
+
 ```
 
 ## Sample Output
 
-```ansi
+```python
 # Language Identification 🌍
 [
   {"lang": "Bemba",  "conf": 0.96},
@@ -175,24 +140,21 @@ print(topic_results)
 ```
 
 
-
-
 ```
-=========================== MULTI‑TASK PIPELINE ===========================
+=========================== Training Architecture ===========================
 
-📥 Input                →          🧠 Core Engine              →        📈 Output
+📥 Input                →  🧠 Core Engine              →            📈 Output
 ------------------------------------------------------------------------------------
-Text (Any Language)     →   Tokenizer 🔤                       →   Language 🌍
+Text (Any Language)     →   Tokenizer 🔤                       →     Language 🌍
                         →   Shared mBERT Encoder 🧠            →     Bemba / Nyanja /
                         →   CLS Vector 🎯                      →     English / Mixed
 ------------------------------------------------------------------------------------
-User Feedback 💬        →   Tokenizer 🔤                       →   Sentiment ❤️
+User Feedback 💬        →   Tokenizer 🔤                       →     Sentiment ❤️
                         →   Shared Encoder 🧠                  →     Negative / Neutral /
                         →   CLS Vector 🎯                      →     Positive
 ------------------------------------------------------------------------------------
-Ride Context 🚗         →   Tokenizer 🔤                       →   Topic 🗂️
+Ride Context 🚗         →   Tokenizer 🔤                       →     Topic 🗂️
                         →   Shared Encoder 🧠                  →     Driver / Payment /
                         →   CLS Vector 🎯                      →     Support / App / Availability
 ------------------------------------------------------------------------------------
-
 ```