andruhon's picture
Update README.md
4ee7dfe verified
|
raw
history blame
815 Bytes
metadata
language:
  - en
license: apache-2.0

Classify text by UNSPSC family

Forked from https://huggingface.co/govspend/unspsc_family_5examples_test2

See https://en.wikipedia.org/wiki/UNSPSC

Usage

There was no documentation on the original model, and demo isn't working, but this works with simple pipeline:

pipe = pipeline("text-classification", model="andruhon/unspsc_family_5examples_test2", tokenizer="bert-base-uncased")
pipe("7oz hammer");
# Would return something like {'label': 'LABEL_105', 'score': 0.339}
# In this case LABEL_105 clearly goes into 27110000 Handtools 

License

The original model didn't have license file. Considering that it's BERT it should have the same license, which I think is Apache 2.0. Use on your own risk. I'll update this file once I have more info.