News-Translation-and-summarization

Comparative study of summarization models - Abstractive and Extractive

Abstractive models:

1. Pegasus
2. T-5

Extractive models:

1. Luhn
2. Lexrank

Steps to run:

1. Load the notebook into Google Colab platform or open using a Jupyter Notebook.
2. Then click "Run all" under runtime ribbon.
3. Enter the German news that you want to translate, when asked.
4. You'll notice that the news has been translated and displayed.
5. You will be asked to enter the reference translation of the same news.
6. You'll be provided with the score of the translation.
7. Following which you'll encounter the summarized news by the two varients of the Summarizers.
8. You can look at the ROUGE scores that follow the above step.
9. You can use this notebook to analyze the different types and models that are best suited for the process of news summarization.
10. Incase the content helped you explore/learn something new; Do give it a star in GitHub.