Skip to content

Latest commit

 

History

History
47 lines (32 loc) · 1.21 KB

README.md

File metadata and controls

47 lines (32 loc) · 1.21 KB
Raw

Psycho Embeddings

A Python package to extract contextualised embeddings. Given a sentence (e.g., the cat is on the table) and a word (e.g., cat) we can extract the embedding of the word cat in the sentence.

Getting Started

We require a python environment with a fully functional PyTorch installation. Then, please install our dependencies with:

git clone https://github.com/MilaNLProc/psycho-embeddings.git
pip install -e psycho-embeddings

Examples

Extract the contextualised embedding of words in context

You can request representations:

  • for one or more layers (layers_id)
  • including static non-contextualised vectors (return_static)
from psycho_embeddings import ContextualizedEmbedder
model = ContextualizedEmbedder("bert-base-cased", max_length=128)

embeddings = model.embed(
    words=["play", "play"],
    target_texts=["I like the way you play.", "The play was outstanding."],
    layers_id=range(13),
    batch_size=8,
    return_static=True,
)

Reference

Credits

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.