https://github.com/amitkumarj441/cikm2023_subspaceembedding

Pluggable Embedding Code for our CIKM paper titled "Lightweight Adaptation of Neural Language Models via Subspace Embedding"
https://github.com/amitkumarj441/cikm2023_subspaceembedding

embeddings language-model nlp

Last synced: 3 months ago
JSON representation

Pluggable Embedding Code for our CIKM paper titled "Lightweight Adaptation of Neural Language Models via Subspace Embedding"

• Host: GitHub
• URL: https://github.com/amitkumarj441/cikm2023_subspaceembedding
• Owner: amitkumarj441
• License: apache-2.0
• Created: 2023-08-16T13:21:37.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2024-05-07T00:03:58.000Z (about 1 year ago)
• Last Synced: 2025-01-26T14:11:58.181Z (5 months ago)
• Topics: embeddings, language-model, nlp
• Language: Python
• Homepage:
• Size: 7.81 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# CIKM2023_SubspaceEmbedding

Pluggable Embedding Code of our CIKM paper titled "Lightweight Adaptation of Neural Language Models via Subspace Embedding"

#### Subspace Embedding

- Embedding layer can be customised.

- Follow the parameters detail in the paper to reproduce the subspace of the embedding.

- Practically, the training/inference speed is similar with Embedding layer.