From a50fa469492e9bbb53dcdfeaa54ea9d2aca8d808 Mon Sep 17 00:00:00 2001 From: Tyriek Warren <62261407+Tyriek-cloud@users.noreply.github.com> Date: Wed, 20 Nov 2024 21:58:35 -0500 Subject: [PATCH] Update README.md Adding an update about the Image Generator. --- README.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index bcb127c..d0902a0 100644 --- a/README.md +++ b/README.md @@ -6,7 +6,7 @@ Wikipedia Article Analyzer: https://wikipedia-article-analyzer-nlp-app-app-ggh9s This project is designed to scrape articles from Wikipedia. By default, the Streamlit application will display the Wikipedia Statistics page (https://en.wikipedia.org/wiki/Statistics). The user will have the option to insert their own Wikipedia URLs. The end goal is to use Natural Language Processing to create a summary (with an emphasis on important and easily digestable details) from Wikipedia articles and display important images/pictures/graphics from each of the articles. -The user can now also interact with an Image Generator. The Image Generator is currently a work-in progress. The Image Generator is being trained by an external model from Hugging Face. The app's author will be continuously training the external model to produce better results. +The user can now also interact with an Image Generator (powered by Stable Diffusion). The Image Generator is an ongoing/work-in progress feature. The app's author will be continuously training the external model and the application to produce better results. # Summary @@ -14,6 +14,8 @@ The application generates a summary of 10+ sentences. The sentences are intended The user may also ask questions about the Wikipedia article that was analyzed. Be sure to hit the "Analyze" button before trying to ask a question. This will ensure that the app webscrapes the relevant Wikipedia article. +The user can now utilize an Image Generator. The idea is that the user can take information that is being summarized in the articles and "visualize the summary" using the image generation feature. It is sometimes useful to look at pictures as opposed to text when learning (or at least that has been the author's experience). + # Image URLs This will pull in links to images from the Wikipedia article. These URLs are clickable and will generate a seperate webpage. The images can then be copied for personal use. @@ -22,7 +24,7 @@ This will pull in links to images from the Wikipedia article. These URLs are cli The structure of the Wikiepedia article may impact whether a summary is actually generated. Image URLs are always pulled in. -The base code for this app was written in Google Colab: https://colab.research.google.com/drive/1RDKGcidcVlpzs2kT-9SB0WKN2NaA4sft?usp=sharing. Mid-project, I realized that the "spacy" library would be compatible on Google Colab but not on Streamlit.io. +The original base code for this app was written in Google Colab: https://colab.research.google.com/drive/1RDKGcidcVlpzs2kT-9SB0WKN2NaA4sft?usp=sharing. Mid-project, I realized that the "spacy" library would be compatible on Google Colab but not on Streamlit.io. Overall, the project continues to evolve and deviate from the base Jupyter Notebook. So I researched solutions to modify the tokenization and other features of the application.