SPACY v3: Custom trainable relation extraction component

spaCy is a popular open-source library for industrial-strength Natural Language Processing in Python. spaCy v3.0 features new transformer-based pipelines that get spaCy’s accuracy right up to the current state-of-the-art, and a new training config and workflow system to help you take projects from prototype to production. In this video, I’ll show you how to apply all these new features when implementing a custom trainable component from scratch.

STEP BY STEP
00:00 – Intro to spaCy v3
01:06 – Overview of the relation extraction challenge
04:02 – Building the ML model: schematic overview
08:07 – Implementation of the ML model in Thinc
18:27 – Defining the configuration file
22:55 – Overview of the TrainablePipe API
25:30 – Using a custom extension attribute
26:20 – Implementation of the custom pipeline component
30:40 – Recap and overview
32:08 – Executing the code as a spaCy project
34:39 – Using a Transformer model
37:16 – Summary and conclusion

SPACY
● Website & documentation: https://spacy.io
● GitHub: https://github.com/explosion/spaCy
● Free online course: https://course.spacy.io
● Thinc: https://thinc.ai
● Prodigy: https://prodi.gy/

THIS VIDEO
● Full code as a spaCy project: https://github.com/explosion/projects...
● What's new in spaCy v3.0: https://spacy.io/usage/v3
● Processing pipelines: https://spacy.io/usage/processing-pip...
● Implementing custom neural networks for spaCy: https://spacy.io/usage/layers-archite...

FOLLOW US
● Sofie Van Landeghem:   / oxykodit  
● spaCy:   / spacy_io  
● Explosion:   / explosion_ai  

CREDITS
● Icons: https://twemoji.twitter.com