https://github.com/veldhub/veld_chain__mara_load_and_publish_models

Chain velds for publishing self-trained MARA models to huggingface.
https://github.com/veldhub/veld_chain__mara_load_and_publish_models

nlp spacy spacy-nlp

Last synced: about 2 months ago
JSON representation

Chain velds for publishing self-trained MARA models to huggingface.

• Host: GitHub
• URL: https://github.com/veldhub/veld_chain__mara_load_and_publish_models
• Owner: veldhub
• License: mit
• Created: 2024-12-04T13:09:42.000Z (5 months ago)
• Default Branch: main
• Last Pushed: 2025-01-20T19:25:33.000Z (4 months ago)
• Last Synced: 2025-01-28T03:56:11.847Z (3 months ago)
• Topics: nlp, spacy, spacy-nlp
• Language: Python
• Homepage:
• Size: 8.79 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# ![veld chain](https://raw.githubusercontent.com/veldhub/.github/refs/heads/main/images/symbol_V_letter.png) veld_chain__mara_load_and_publish_models

**Experimental Work in progress!**

A chain veld repo for publishing self-trained MARA models to huggingface.

These spacy models were trained with spacy 2.2.4 . But only from 3.* onwards, does 

spacy offer a native huggingface integration which converts model metadata into harmonized 

huggingface metadata. Spacy 2.* does not offer this and it also does not package into whl formats 

which are required by huggingface. 

Further work must be invested into bridging this gap, potentially by manually building a whl file 

with `build setuptools wheel` and a `pyproject.toml` file.