πŸ† #2,207 overall#524 of 601 in Language Models

guolinke /TUPE

Transformer with Untied Positional Encoding (TUPE). Code of paper "Rethinking Positional Encoding in Language Pre-training". Improve existing models like BERT.

$ git clone https://github.com/guolinke/TUPE.git
GitHub social preview for guolinke/TUPE
Stars
250
250
Forks
27
27
Language
Python
License
MIT
Created
Jun 24, 2020
6.3 years old
Last push
Nov 8, 2021

Categories

GitHub topics

More in Language Models

Language Models

codertimo/BERT-pytorch

Google AI 2018 BERT pytorch implementation

🧠 Natural Language Processing
PythonApache-2.0updated Sep 15, 2023
GitHub β†—β˜… 6.5Kβ‘‚ 1.3K
Language Models

michiyasunaga/LinkBERT

[ACL 2022] LinkBERT: A Knowledgeable Language Model 😎 Pretrained with Document Links

❓ Question Answering
PythonApache-2.0updated Apr 5, 2022
GitHub β†—β˜… 462β‘‚ 43
Language Models

michiyasunaga/dragon

[NeurIPS 2022] DRAGON 🐲: Deep Bidirectional Language-Knowledge Graph Pretraining

❓ Question Answering
PythonApache-2.0updated May 10, 2023
GitHub β†—β˜… 336β‘‚ 46
Language Models

extreme-bert/extreme-bert

ExtremeBERT is a toolkit that accelerates the pretraining of customized language models on customized datasets, described in the paper β€œExtremeBERT: A Toolkit for Accelerating Pretraining of Customized BERT”.

🧠 Natural Language Processing
PythonApache-2.0updated Mar 5, 2023
GitHub β†—β˜… 267β‘‚ 14

Data from GitHub Β· snapshot Sep 24, 2026