Update README.md

README.md

@@ -27,6 +27,8 @@ pinned: false
 
 ![Frame 1](https://user-images.githubusercontent.com/65676392/227686139-04784d02-c1b0-4911-a0d4-e878a954e5c2.png)
 
+<br>
+
 # Methodology
 
 BERT (Bidirectional Encoder Representation from Transformers) is a pre-trained language model designed to improve nautral-language understanding in various tasks. BERT can find the answer given a question and reference by learning contextual relationships between words in a a text using bidirectional transformers. 
@@ -39,7 +41,7 @@ We fine-tuned BERT for question answering using the Stanford Question Answering
 4. **Answer prediction**: BERT generates a probability distribution over the tokens in the reference text for both the start and end positions of the answer. The model identifies the tokens with the highest probabilities for the start and end positions as the answer span.
 5. **Answer extraction**: The answer is extracted by combining the tokens from the identified start to end positions.
    
-   
+  <br>
 # References
 [1] https://huggingface.co/ml6team/keyphrase-extraction-kbir-inspec <br>
 [2] https://arxiv.org/pdf/1810.04805v2.pdf <br>