Update README.md

README.md

@@ -95,6 +95,8 @@ for text, result in zip(texts, results):
 
 ## 📊 Performance Overview
 
+![RoBERTa-Sentimentic Performance](roberta_sentimentic_performance.png)
+
 ### Benchmark Results
 
 | Dataset | Pre-trained RoBERTa | RoBERTa-Sentimentic | Improvement |
@@ -113,6 +115,8 @@ for text, result in zip(texts, results):
 
 ## 🎯 Model Performance Analysis
 
+![Cross-Domain Transfer](cross_domain_transfer.png)
+
 ### Confusion Matrices
 
 #### IMDB Dataset Results
@@ -146,6 +150,8 @@ Precision: 91.5% | Recall: 91.4% | F1-Score: 91.5%
 
 ## 🛠️ Technical Details
 
+![Model Architecture](model_architecture.png)
+
 ### Architecture
 - **Base Model**: [roberta-base](https://huggingface.co/roberta-base) (125M parameters)
 - **Task Head**: Linear classification layer with dropout (0.1)
@@ -183,6 +189,8 @@ Training Time: ~25 minutes total
 
 ## 📈 Training Process
 
+![Training Progress](training_progress.png)
+
 ### Phase 1: IMDB Fine-tuning
 - **Dataset**: 25,000 IMDB movie reviews  
 - **Strategy**: Same-domain fine-tuning