Text Summarization with Google AI's T5 in Python

preview_player
Показать описание
Easy text summarization using Google AI's T5 model using HuggingFace transformers and PyTorch in Python.

Automatic text summarization allows us to shorten long pieces of text into easy-to-read, short snippets that still convey the most important and relevant information of the original text.

In this video, we’ll build a simple but incredibly powerful text summarizer using Google’s T5. We’ll be using the PyTorch and HuggingFace’s Transformers frameworks.

This is split into three parts:
1. Import and Initialization
2. Data and Tokenization
3. Summary Generation

🤖 70% Discount on the NLP With Transformers in Python course:

You can read the article version of this on Medium here:

(And for those of you without Medium membership, here's a free link):

Thumbnail background by gustavo centurion on Unsplash
Рекомендации по теме
Комментарии
Автор

Found your youtube in medium article, you’re doing amazing job, keep posting!! Also, a suggestion: there’s not a lot of production ml content on youtube.

pauliuspuskunigis
Автор

Great job, thanks a lot for this tutorial !

Dmitriuso
Автор

Hi James, do you recommend this model for finding entries in a database that are duplicates semantically speaking? I'm looking for a model to do that, to catch if i'm inserting something to a database that is was already there but just in different words...

mihaelacostea
Автор

Can you tell me how to finetune model with long text inputs?. I think if set a fixed max length, the input will loss information

ngochuannguyen
Автор

Good job man! I really liked this. THUMBS UP

mohamadrezabidgoli
Автор

What's the longest amount of text you can enter in? Can you enter an entire book's worth of text and get a series of summaries back which are up to 512 tokens each?

OccultDemonCassette
Автор

Is this abstractive or extractive summarization? Btw, great video. You really made it look like piece of cake.

siddhantrai
Автор

Hi James, to work with other languages can I use the TM5 model ? Do I need to change the code or just the model name?

muzaffarismatillayev
Автор

Hi James, what do I need to do in order to make the model works with other languages? What’s more, is there a way for fine-tuning this model? Many thanks!

jarorkwong
Автор

your text is e, where did you use it in coming lines? I somehow missed it.

sayajujur
Автор

Librarys are not worked with windows 🤔

smtsdlkr