TsinghuaAI / TDSLinks
A plug-in of Microsoft DeepSpeed to fix the bug of DeepSpeed pipeline
☆26Updated 4 years ago
Alternatives and similar repositories for TDS
Users that are interested in TDS are comparing it to the libraries listed below
Sorting:
- Pretrain CPM-1☆51Updated 4 years ago
- Code for the paper "A Theoretical Analysis of the Repetition Problem in Text Generation" in AAAI 2021.☆54Updated 2 years ago
- ☆45Updated 3 years ago
- Code for ACL 2023 paper titled "Lifting the Curse of Capacity Gap in Distilling Language Models"☆28Updated last year
- Notes of my introduction about NLP in Fudan University☆37Updated 3 years ago
- A pre-trained model with multi-exit transformer architecture.☆54Updated 2 years ago
- Inference framework for MoE layers based on TensorRT with Python binding☆41Updated 4 years ago
- Ongoing research training transformer language models at scale, including: BERT & GPT-2☆69Updated last year
- ☆53Updated 3 years ago
- BANG is a new pretraining model to Bridge the gap between Autoregressive (AR) and Non-autoregressive (NAR) Generation. AR and NAR generat…☆28Updated 3 years ago
- EMNLP 2021: Single-dataset Experts for Multi-dataset Question-Answering☆70Updated 3 years ago
- Method to improve inference time for BERT. This is an implementation of the paper titled "PoWER-BERT: Accelerating BERT Inference via Pro…☆61Updated last month
- Source code for paper: Knowledge Inheritance for Pre-trained Language Models☆38Updated 3 years ago
- reStructured Pre-training☆98Updated 2 years ago
- Finetune CPM-1☆74Updated 2 years ago
- Princeton NLP's pre-training library based on fairseq with DeepSpeed kernel integration 🚃☆114Updated 2 years ago
- Paradigm shift in natural language processing☆42Updated 3 years ago
- Must-read papers on improving efficiency for pre-trained language models.☆104Updated 2 years ago
- Finetune CPM-1☆24Updated 4 years ago
- This repository is the official implementation of our EMNLP 2022 paper ELMER: A Non-Autoregressive Pre-trained Language Model for Efficie…☆26Updated 2 years ago
- ☆105Updated 2 years ago
- Source code for NAACL 2021 paper "TR-BERT: Dynamic Token Reduction for Accelerating BERT Inference"☆47Updated 3 years ago
- This package implements THOR: Transformer with Stochastic Experts.☆65Updated 3 years ago
- ☆24Updated 2 years ago
- Retrieval as Attention☆82Updated 2 years ago
- ☆67Updated 3 years ago
- ☆38Updated 11 months ago
- Pytorch implementation of paper "Efficient Nearest Neighbor Language Models" (EMNLP 2021)☆73Updated 3 years ago
- [NAACL 2021] Factual Probing Is [MASK]: Learning vs. Learning to Recall https://arxiv.org/abs/2104.05240☆168Updated 2 years ago
- ☆17Updated 5 years ago