Quantization made by Richard Erkhov.

[Github](https://github.com/RichardErkhov)

[Discord](https://discord.gg/pvy7H8DZMG)

[Request more models](https://github.com/RichardErkhov/quant_request)


Pretrain-Qwen-1.2B - GGUF
- Model creator: https://huggingface.co/MiniLLM/
- Original model: https://huggingface.co/MiniLLM/Pretrain-Qwen-1.2B/


| Name | Quant method | Size |
| ---- | ---- | ---- |
| [Pretrain-Qwen-1.2B.Q2_K.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q2_K.gguf) | Q2_K | 0.51GB |
| [Pretrain-Qwen-1.2B.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q3_K_S.gguf) | Q3_K_S | 0.57GB |
| [Pretrain-Qwen-1.2B.Q3_K.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q3_K.gguf) | Q3_K | 0.61GB |
| [Pretrain-Qwen-1.2B.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q3_K_M.gguf) | Q3_K_M | 0.61GB |
| [Pretrain-Qwen-1.2B.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q3_K_L.gguf) | Q3_K_L | 0.63GB |
| [Pretrain-Qwen-1.2B.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.IQ4_XS.gguf) | IQ4_XS | 0.65GB |
| [Pretrain-Qwen-1.2B.Q4_0.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q4_0.gguf) | Q4_0 | 0.67GB |
| [Pretrain-Qwen-1.2B.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.IQ4_NL.gguf) | IQ4_NL | 0.67GB |
| [Pretrain-Qwen-1.2B.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q4_K_S.gguf) | Q4_K_S | 0.69GB |
| [Pretrain-Qwen-1.2B.Q4_K.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q4_K.gguf) | Q4_K | 0.72GB |
| [Pretrain-Qwen-1.2B.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q4_K_M.gguf) | Q4_K_M | 0.72GB |
| [Pretrain-Qwen-1.2B.Q4_1.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q4_1.gguf) | Q4_1 | 0.72GB |
| [Pretrain-Qwen-1.2B.Q5_0.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q5_0.gguf) | Q5_0 | 0.78GB |
| [Pretrain-Qwen-1.2B.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q5_K_S.gguf) | Q5_K_S | 0.79GB |
| [Pretrain-Qwen-1.2B.Q5_K.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q5_K.gguf) | Q5_K | 0.81GB |
| [Pretrain-Qwen-1.2B.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q5_K_M.gguf) | Q5_K_M | 0.81GB |
| [Pretrain-Qwen-1.2B.Q5_1.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q5_1.gguf) | Q5_1 | 0.83GB |
| [Pretrain-Qwen-1.2B.Q6_K.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q6_K.gguf) | Q6_K | 0.93GB |
| [Pretrain-Qwen-1.2B.Q8_0.gguf](https://huggingface.co/RichardErkhov/MiniLLM_-_Pretrain-Qwen-1.2B-gguf/blob/main/Pretrain-Qwen-1.2B.Q8_0.gguf) | Q8_0 | 1.15GB |




Original model description:
---
library_name: transformers
license: apache-2.0
datasets:
- monology/pile-uncopyrighted
- MiniLLM/pile-tokenized
language:
- en
metrics:
- accuracy
pipeline_tag: text-generation
---

# Pretrain-Qwen-1.2B

[paper](https://arxiv.org/abs/2410.17215) | [code](https://github.com/thu-coai/MiniPLM)

**Pretrain-Qwen-1.2B** is a 1.2B model with Qwen achitecture conventionally pre-trained from scratch on [the Pile](https://huggingface.co/datasets/monology/pile-uncopyrighted) for 50B tokens.

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-1.2B](https://huggingface.co/MiniLLM/MiniPLM-Qwen-1.2B)**

## 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
+ [VanillaKD](https://huggingface.co/MiniLLM/VanillaKD-Pretrain-Qwen-1.2B)

## 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}
}
```