wj-mcat / paddle-prompt

Prompt Toolkit for PaddlePaddle Deep Learning Framework

paddle-prompt

Prompt Toolkit for PaddlePaddle Deep Learning Framework. Inspired by :

• thunlp/OpenPrompt
• AutoPrompt

---

Installation

paddle-prompt is built under Python3.7+ and Paddle. You can install paddle-prompt with pip:

pip install paddle-prompt

TODOS

• complete PLMs
• complete template
• complete verbalize
• create text classification datasets
• ...

论文复现

在接下来一段时间内，使用本工具可复现以下论文，如果大家想要复现其它论文，欢迎发Issue和Contribution。

• PET: Exploiting Cloze Questions for Few Shot Text Classification and Natural Language Inference
• P-Tuning: GPT Understands, Too
• P-Tuning v2: Prompt Tuning Can Be Comparable to Fine-tuning Universally Across Scales and Tasks
• LM-BFF: Better Few-shot Fine-tuning of Language Models
• Entailment as Few-Shot Learner
• ...... more contribution

Stargazers over time

Contributors

Made with contrib.rocks

Creators

• @wj-Mcat - Jingjing WU (吴京京)

Copyright & License

• Code & Docs © 2022 wj-Mcat https://github.com/wj-Mcat
• Code released under the Apache-2.0 License
• Docs released under Creative Commons

paddle-prompt's Issues

Implement P-Tuning

paper AdaPrompt: Adaptive Model Training for Prompt-based NLP

Implement Soft Prompt

related papers:

• Prefix-Tuning: Optimizing Continuous Prompts for Generation
• The Power of Scale for Parameter-Efficient Prompt Tuning
• Learning How to Ask: Querying LMs with Mixtures of Soft Prompts

Implement OptimPrompt

paper: Factual Probing Is [MASK]: Learning vs. Learning to Recall

Implement TransPrompt

paper: TransPrompt: Towards an Automatic Transferable Prompting Framework for Few-shot Text Classification

Implement AdaPrompt

paper AdaPrompt: Adaptive Model Training for Prompt-based NLP

Implement feature of paper: LM-BFF

Introduce

• paper
• github code