Here's the link to try it out Live : (https://huggingface.co/spaces/DarkInventor/Summarizer)
As a machine learning enthusiast, I have been immersing myself in exploring the world of pre-existing models. Recently, my excitement peaked when I learned about Google's innovative model, Bart, during their I/O event. However, as a Canadian resident, I was disheartened to discover that Bart cannot be launched in my country due to regulatory and privacy concerns.
Despite this setback, I was determined to try out Bart's capabilities, and so I sought out a solution for those of us who cannot access its API in Canada, as well as in other countries where it is not available. After some research and experimentation, I have successfully built a summarizer app that utilizes the pre-trained "bart-large-cnn" model.
To make the app user-friendly, I utilized Gradio to create a simple and intuitive GUI. Additionally, I imported the "transformers" module from the pipeline API to download and utilize the Bart model. Now, Canadians and those in other restricted countries can enjoy the benefits of Bart's powerful summarization capabilities.
Here, I am entering 1000 words and selecting the Max_length. Length is counted in the form of nltk tokens.