https://github.com/thunlp/KnowledgeablePromptTuning

kpt code
https://github.com/thunlp/KnowledgeablePromptTuning

Last synced: 3 months ago
JSON representation

kpt code

• Host: GitHub
• URL: https://github.com/thunlp/KnowledgeablePromptTuning
• Owner: thunlp
• Created: 2021-09-13T06:33:57.000Z (about 3 years ago)
• Default Branch: main
• Last Pushed: 2023-04-16T16:20:23.000Z (over 1 year ago)
• Last Synced: 2024-06-14T01:46:12.656Z (5 months ago)
• Language: Python
• Size: 211 KB
• Stars: 205
• Watchers: 4
• Forks: 22
• Open Issues: 10
• Metadata Files:
  • Readme: readme.md

Awesome Lists containing this project

• StarryDivineSky - thunlp/KnowledgeablePromptTuning

README

        

# KPT source code

Here is the source code for our ACL 2022 paper 

[Knowledgeable Prompt-tuning: Incorporating Knowledge into Prompt Verbalizer for Text Classification](https://arxiv.org/abs/2108.02035)

## install openprompt 

Please install via git clone. This helps keep the dataset downloading scripts.

```bash

git clone [email protected]:thunlp/OpenPrompt.git

cd OpenPrompt

python setup.py install

```

## Download the dataset

```

cd OpenPrompt/datasets

bash download_text_classification.sh

```

## Run the scripts

for fewshot experiment

```

bash scripts/run_fewshot.sh 

```

for zeroshot experiment

```

bash scripts/run_zeroshot.sh

```

for pilot experiment in appendix

```

bash scripts/run_pilot.sh

```

The possible arguments in the scripts are in the comment of the scripts. 

Please choose the combination according to your need.