Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
| Date | Stars |
|---|---|
| 2026-07-31 | 600 |
| 2026-08-04 | 600 |
| 2026-08-06 | 600 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
## Parameter efficient instruction tuning: an Empirical Study
This repository serves as an effort to systematically to compare different parameter efficient fine-tuning methods on instruction tuning task. We use the [NI dataset](https://github.com/allenai/natural-instructions) as the benchmark dataset. The technical report can be found [here](https://arxiv.org/pdf/2411.16775)
PEFT method implementations are adapted from [adapter-transformers](https://github.com/Adapter-Hub/adapter-transformers) and [peft](https://github.com/huggingface/peft).
## Training
* For bash scripts to run all experiments, refer to `scripts` folder.
* All experiments calling scripts are formatted by `scripts/hfai/hp_run.sh`. For example, to run a LoRA experimental in the development mode, run the following command.
```bash
bash scripts/hfai/hp_run.sh lora_peft dev
```
## Dataset
* We employ [SuperNI](https://github.com/allenai/natural-instructions) as our training and evaluation datasets.
## Setup
To install the package, do the following.
* `conda create -n peft python=3.8`
* `git clone https://github.com/hepengfe/peft-private.git` and checkout `git checkout release-v0.4.0-adapter` branch. Under `peft-private` folder, `pip install -e .` to install peft-private.
<!-- * under `adapater-transfeomers` folder, `pip install -e .` to install adapter-transformer.
* `pip uninstall transfomers` uninstall the original transformer installed by `peft` to use adapter-transformer instead. -->
* rouge-score: `pip install rouge-score`
* make sure GPT2 is under `cache/saved_pretrained/gpt2` for evaluation
## HPC Platform specific
The platform we use is [hfai HPC](https://www.high-flyer.cn/en/). Each node is equipped with A100x8 GPU, and each of our experiments is runing on a single node.
This codebase is highly optimized for hfai platform, and it supports the following functionalities:
* **Experiment Configuration and Submission**: The `hp_run.sh` scripts allows for flexible adjustments to experiment name, batch size and training framework based on fine-tuning methods. To launch multiple jobs based on `hp_run.sh`, refer to scripts under `scripts/hfai` folder.
* **Checkpoint Management**: Since the platform is pre-emptable, our codebase supports checkpoints saving upon suspension and resuming from the last checkpoint. Each experiment is assumed to be run until complete test dataset evaluation.
* **Training State Validation**: When saving checkpoint, we support checking the completeness of training state, training random states. Otherwise, it will delete the latest checkpoint and needs re-run the experiment loading second-to-last checkpoint.
* **System Message and Debugging**: We let most system message output by print statement because it's more suitable for multi-process debugging, and we suppress warnings that contain `error` string to avoid job killing.
Here are some extra notes for hfai platform:
* default ni dataset dir is `../../data` due to hfai compatibility.
* **Pytorch and CUDA compatibility**: hfai platform has CUDA version 11.3, and [peft](https://github.com/huggingface/peft) setup requires `torch>=1.13.0`. Therefore, we use the corresponding pytorch version `1.10.2+cu113` by peft setup.
## Citation
If you found the codebase or my work valuable, please cite:
```
@misc{he2024parameterefficientinstructiontuning,
title={Parameter Efficient Instruction Tuning: An Empirical Study},
author={Pengfei He},
year={2024},
eprint={2411.16775},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2411.16775},
}
```
Excerpt of 3,613 characters
Read on GitHub
Sourab Mangrulkar · Amazon · India
191
97
Younes B
17
Sayak Paul · @huggingface · India
14
Saurav Maheshkar · United Kingdom
7
Mayank Mishra · UC Berkeley
3
Kashif Rasul · Germany
2
2
Lysandre Debut · Hugging Face · France
1
Zach Nussbaum
1
muhtasham · stealth · United States
1
1
Manuel Romero · ClibrAIn · Spain
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:bf89cd50c0102045, name:parameter-efficient, name:instruction tuning