Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/pandermatt/simpletext-clef2023-2024

Last synced: 6 days ago
JSON representation

Host: GitHub
URL: https://github.com/pandermatt/simpletext-clef2023-2024
Owner: pandermatt
Created: 2024-06-21T17:03:08.000Z (5 months ago)
Default Branch: main
Last Pushed: 2024-06-21T17:07:42.000Z (5 months ago)
Last Synced: 2024-06-22T09:36:58.217Z (5 months ago)
Size: 1000 Bytes
Stars: 0
Watchers: 1
Forks: 0
Open Issues: 0
Metadata Files:
- Readme: README.md

Awesome Lists containing this project

README

        # SimpleText Best of Labs in CLEF-2023: Scientific Text Simplification Using Multi-Prompt Minimum Bayes Risk Decoding

This repository includes a Jupyter Notebook (`clef2024-summary.ipynb`) that accompanies the paper. The notebook is designed to replicate the experiments discussed in the paper, particularly focusing on text simplification using multiple prompt strategies and Minimum Bayes Risk (MBR) decoding.

> [!NOTE]

> Note that we haven’t published the entire codebase for the experiments in the paper. If you have any questions or need further information, please contact us.

## Resources

- **Llama 3 Model:** The simplifications in this study are generated using the Llama 3 model. You can find more information about the model here: [Meta Llama 3 - 8B Instruct on Hugging Face](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct).

- **LENS Metric:** The evaluation metric used in the Minimum Bayes Risk decoding process is LENS. More details about LENS can be found in the official GitHub repository: [LENS GitHub Repository](https://github.com/Yao-Dou/LENS).

## Code from the Previous Paper (CLEF-2023)

The folder [prompts-2023](prompts-2023) contains the prompts used for the previous paper "UZH_Pandas at SimpleText@CLEF-2023: Alpaca LoRA 7B and LENS Model Selection for Scientific Literature Simplification". Note that the code for the fine-tuning of Alpaca LoRA 7B can be found in the [Alpaca LoRA 7B repository](https://github.com/tloen/alpaca-lora).

  Fine-tuning Parameters

  ```python

  train(

      base_model="chainyo/alpaca-lora-7b",

      data_path=filename,

      prompt_template_name=template,

      num_epochs=3,

      cutoff_len=512,

      batch_size=64,

      group_by_length=True,

      val_set_size=0.2,

      output_dir=config.data_dir(f"alpaca-lora-both-{template}"),

      lora_target_modules=["q_proj", "k_proj", "v_proj", "o_proj"],

      lora_r=16,

      micro_batch_size=8,

  )

```

## Contributors

- [Andrianos Michail](https://www.cl.uzh.ch/de/about-us/people/team/compling/amichail.html) ([email protected])

- [Pascal Severin Andermatt](https://www.ifi.uzh.ch/en/ddis/people/pandermatt.html) ([email protected])

- Tobias Fankhauser ([email protected])