|
--- |
|
library_name: transformers |
|
license: apache-2.0 |
|
datasets: |
|
- monology/pile-uncopyrighted |
|
- MiniLLM/pile-tokenized |
|
language: |
|
- en |
|
metrics: |
|
- accuracy |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
# VanillaKD-Pretrain-Qwen-200M |
|
|
|
[paper](https://arxiv.org/abs/2410.17215) | [code](https://github.com/thu-coai/MiniPLM) |
|
|
|
**VanillaKD-Pretrain-Qwen-200M** is a 200M model with Qwen achitecture pre-trained with vanilla token-level knowledge distillation on [the Pile](https://huggingface.co/datasets/monology/pile-uncopyrighted) for 50B tokens. The teacher model is [Qwen1.5-1.8B](https://huggingface.co/Qwen/Qwen1.5-1.8B). |
|
|
|
We also open-source the tokenized [pre-training corpus](https://huggingface.co/datasets/MiniLLM/pile-tokenized) for reproducibility. |
|
|
|
**It is used as the baseline for [MiniLLM-Qwen-200M](https://huggingface.co/MiniLLM/MiniPLM-Qwen-200M)** |
|
|
|
## Evaluation |
|
|
|
MiniPLM models achieves better performance given the same computation and scales well across model sizes: |
|
|
|
<p align='left'> |
|
<img src="https://cdn-uploads.huggingface.co/production/uploads/624ac662102fcdff87be51b9/EOYzajQcwQFT5PobqL3j0.png" width="1000"> |
|
</p> |
|
|
|
## Other Baselines |
|
+ [Conventional Pre-Training](https://huggingface.co/MiniLLM/Pretrain-Qwen-200M) |
|
|
|
## Citation |
|
|
|
```bibtext |
|
@article{miniplm, |
|
title={MiniPLM: Knowledge Distillation for Pre-Training Language Models}, |
|
author={Yuxian Gu and Hao Zhou and Fandong Meng and Jie Zhou and Minlie Huang}, |
|
journal={arXiv preprint arXiv:2410.17215}, |
|
year={2024} |
|
} |
|
``` |
