code gist

Team-5
 Problem
statement
To explore the development of a
sophisticated code summarization tool to
enhance code readability, improve
software documentation quality, and
accelerate code comprehension and
collaboration among developers.
 Why is the proposed solution
better than the existing ones?

Deep learning Approach:leveraging transformer-based

models,pre-trained on code and our tool generates that
summary
Fine-tuning and optimization:by fine-tuning of pre-trained
transformer model our tool accurately generate summaries
Efficient data preparation:Our tool utilizes advanced
tokenization techniques for code ensuring appropriately
processing of model
 Technical
work details
Pre-processing the data: parseing
cleaning and tokenizing the code and
constructing a tree representation
LSTM layers: Key component of the
project. The pipeline used includes
Child-Sum Tree LSTM, N-ary LSTM
and Bidirectional LSTM
Contributions:

CS21B014: testing, documentation,preprocessing and model training

implementation
CS21B016: Data collection,testing and documentation
CS21B019: Documentation, Preprocessing and model training
CS21B021: Documentation, model training implementation and
testing
CS21B034: Data collection, testing and documentation
Plan for release-2

Model improvement
Integration with vs code
Dataset expansion
Dynamic working
THANK YOU