https://github.com/epfml/x2static
X2Static embeddings
https://github.com/epfml/x2static
Last synced: 10 months ago
JSON representation
X2Static embeddings
- Host: GitHub
- URL: https://github.com/epfml/x2static
- Owner: epfml
- License: apache-2.0
- Created: 2021-05-31T09:32:30.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2021-07-15T12:26:09.000Z (over 4 years ago)
- Last Synced: 2025-04-18T16:16:09.212Z (11 months ago)
- Language: Python
- Size: 21.5 KB
- Stars: 14
- Watchers: 4
- Forks: 3
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# X2Static vectors
This folder contains the code for training X2Static models.
## Code
The folder src contains the code and instructions to train and reproduce the models trained for our experiments.
## Evaluation
Our models are evaluated using the standard evaluation tool in the [MUSE](https://github.com/facebookresearch/MUSE) repository by Facebook AI Research.
## Pretrained vectors
Pretrained vectors in .bin format can be download from [here](https://zenodo.org/record/5055755). They can be converted to standard word2vec format using the script provided in the src folder. For example, to convert "X2Static_best.bin" to "X2Static_best.vec", use
```
python convert_bin_to_w2v.py X2Static_best.bin X2Static_best.vec
```
## Dataset
Dataset in the required format for the code can be download from [here](https://zenodo.org/record/5055755).
## References
When using this code or some of our pretrained vectors for your application, please cite the following paper:
Prakhar Gupta, Martin Jaggi, [*Obtaining Better Static Word Embeddings Using Contextual Embedding Models*](https://arxiv.org/abs/2106.04302) ACL 2021
```
@inproceedings{Gupta2021ObtainingPC,
title={Obtaining Better Static Word Embeddings Using Contextual Embedding Models},
author={Prakhar Gupta and Martin Jaggi},
booktitle={ACL},
year={2021}
}
```