Using GPT-3 to Generate Answers From Book

HuggingFace Space: https://huggingface.co/spaces/pritish/BookGPT
Kaggle Notebook: https://bit.ly/3kg3dP7

Have you ever wished you could search a book like you search the web? In this video, we'll show you how to build an AI book search using the Universal Sentence Encoder and GPT-3. We'll walk through the steps of breaking down a large book into smaller chunks, generating embeddings for each chunk using the Universal Sentence Encoder, and then using K-nearest neighbors to find the most similar chunks to a user's question. With this approach, you can quickly search through a large corpus of text and find the most relevant information. Finally, we will use GPT3 to generate a concise answer which can also refer page numbers from the book.

Movie Recommendation System: https://youtu.be/pvY0BmAFxwg
(watch this for more info on universal sentence encoder)

🔗 Social Media 🔗
📱 Twitter: https://bit.ly/42t69cG
📝 LinkedIn: https://bit.ly/3JxvkSJ
📂 GitHub: https://bit.ly/3lye8UV

gpt 3 , natural language processing, machine learning, deep learning, Universal Sentence Encoder, K Nearest Neighbor algorithm, embeddings, semantic search, text analysis, data science, artificial intelligence, Python, recommendation systems, HuggingFace

Google Search Video used from 0:19 to 0:21 and 2:04 to 2:12 -
https://youtu.be/0eKVizvYSUQ

Music from Uppbeat (free for Creators!):
https://uppbeat.io/t/jonny-boyle/funny-in-france
License code: KC0DEJS22SOOZIDE

Music from Uppbeat (free for Creators!):
https://uppbeat.io/t/soundroll/that-vibe
License code: ZT1NA5TMCOVFGSRJ

Thank you,
Pritish Mishra