Update README.md

README.md

@@ -113,9 +113,9 @@ The API returns a tuple with two elements:
 
 # Chosen LLM and Justification
 
-We have chosen **DistilBERT** as the foundational LLM for text classification due to its efficiency, lightweight architecture, and high performance in natural language processing (NLP) tasks. DistilBERT is a distilled version of BERT that retains 97% of BERT’s performance while being 60% faster and requiring significantly fewer computational resources. This makes it ideal for classifying extracted text as spam or not spam in real-time OCR applications.
+I have chosen **DistilBERT** as the foundational LLM for text classification due to its efficiency, lightweight architecture, and high performance in natural language processing (NLP) tasks. DistilBERT is a distilled version of BERT that retains 97% of BERT’s performance while being 60% faster and requiring significantly fewer computational resources. This makes it ideal for classifying extracted text as spam or not spam in real-time OCR applications.
+[reference](https://arxiv.org/pdf/1910.01108)
 
----
 
 ## Steps for Fine-Tuning or Prompt Engineering
 
@@ -131,12 +131,6 @@ We have chosen **DistilBERT** as the foundational LLM for text classification du
 4. Implement cross-entropy loss and optimize with AdamW.
 5. Evaluate performance using precision, recall, and F1-score.
 
-### Prompt Engineering (Alternative Approach):
-- If fine-tuning is not preferred, use predefined prompts with a larger LLM (e.g., GPT) to classify text dynamically.
-- Example prompt:  
-
-
----
 
 ## Integration with OCR Output
 
@@ -144,7 +138,6 @@ We have chosen **DistilBERT** as the foundational LLM for text classification du
 - The classification result is appended to the OCR output and stored in `ocr_results.json` and `ocr_results.csv`.
 - The system updates the UI in real-time via **Gradio** to display extracted text along with the classification label.
 
----
 
 ## Security and Evaluation Strategies
 
@@ -156,10 +149,4 @@ We have chosen **DistilBERT** as the foundational LLM for text classification du
 ### Evaluation Strategies:
 - Perform cross-validation to assess model robustness.
 - Continuously monitor classification accuracy on new incoming data.
-- Implement feedback mechanisms for users to report misclassifications and improve the model.
-
----
-
-This integration of OCR and LLM ensures an efficient, scalable, and accurate system for spam classification of text extracted from images.
-
-
+- Implement feedback mechanisms for users to report misclassifications and improve the model.