Forge-Qwen2.5-0.5B-s1
This model is initialized from the pre-trained Qwen/Qwen2.5-0.5B and fine-tuned on the AL-GR/AL-GR-v1 dataset using the FORGE framework for 1 epoch.
It is presented in the paper FORGE: Forming Semantic Identifiers for Generative Retrieval in Industrial Datasets.
- GitHub Repository: https://github.com/selous123/al_sid
- Project Page: https://huggingface.co/AL-GR
Evaluation Results on AL-GR/AL-GR-v1
| Model | HR@20 | HR@100 | HR@500 | HR@1000 |
|---|---|---|---|---|
| Forge-Qwen 2.5-0.5B-Base-s1 | 0.0506 | 0.1277 | 0.2602 | 0.3068 |
| Forge-T5-Base-s1 | 0.0284 | 0.0689 | 0.1372 | 0.1557 |
Note: HR@K denotes Hit Rate at K — the proportion of test queries for which the correct answer appears in the top-K retrieved/generated results.
Usage
1. Download the Model
You can download this model locally using the huggingface_hub library:
import os
os.environ["HF_ENDPOINT"] = "https://hf-mirror.com" # Optional: use mirror for faster download in some regions
os.environ["KMP_DUPLICATE_LIB_OK"] = "True"
from huggingface_hub import snapshot_download
snapshot_download(
repo_id='AL-GR/Forge-Qwen2.5-0.5B-s1',
local_dir='{YOUR_LOCAL_DIR}', # Replace with your desired local path
local_dir_use_symlinks=False,
)
For more details about the training setup, dataset, or evaluation protocol, please refer to the FORGE framework repository.
Citation
@article{fu2025forge,
title={FORGE: Forming Semantic Identifiers for Generative Retrieval in Industrial Datasets},
author={Fu, Kairui and Zhang, Tao and Xiao, Shuwen and Wang, Ziyang and Zhang, Xinming and Zhang, Chenchi and Yan, Yuliang and Zheng, Junjun and others},
journal={arXiv preprint arXiv:2509.20904},
year={2025}
}