ppolx's picture
Models
38a0b29
metadata
license: apache-2.0
language:
  - de
base_model:
  - spacy/de_core_news_sm
library_name: spacy
tags:
  - e-portfolios
  - computer-science
  - learning

AISOP-MdL-classifiers

This is a series of spacy models for the classification tasks.

Try it here

  • Install spaCy
  • python python recognize.py l1-model l2-models "Outro intro zooming"

Yields an output such as {"Tonebene": {"score": 0.99, "subs": {"Musik": 0.4, "Geraeusche": 0.34, "Sprache": 0.3}}}

Web-App Packaging

This model is part of the AISOP-domain-fundid https://gitlab.com/aisop/aisop-domain-mdl which is designed to serve for the AISOP-webapp.

The python scripts there use the models and the spaCy library to classify each "paragraph" of e-portfolios stored in HTML and even generate words using tesseract (if the picture is available) and annotate them too.

The scripts enrich the HTML with data-topic-* attributes, indicating the presence of topics in the paragraphs.

The scripts can be tested in the web-app in the /debug/ road.