|
|
--- |
|
|
license: other |
|
|
license_name: tongyi-qianwen |
|
|
license_link: https://huggingface.co/Qwen/Qwen1.5-7B-Chat/blob/main/LICENSE |
|
|
datasets: |
|
|
- LooksJuicy/ruozhiba |
|
|
- TigerResearch/sft_zh |
|
|
- silk-road/alpaca-data-gpt4-chinese |
|
|
language: |
|
|
- zh |
|
|
- en |
|
|
tags: |
|
|
- Transformer |
|
|
- text-generation-inference |
|
|
--- |
|
|
|
|
|
### Coming Soon!!!!!! |
|
|
|
|
|
### 使用数据集alpaca-data-gpt4-chinese、sft_zh、ruozhiba对Qwen1.5-7B-Chat微调，测试结果显示CEVAL和MMLU分数均有上升 |
|
|
|
|
|
### 模型： |
|
|
- https://huggingface.co/Qwen/Qwen1.5-7B-Chat |
|
|
|
|
|
### 数据集： |
|
|
- https://huggingface.co/datasets/TigerResearch/sft_zh |
|
|
- https://huggingface.co/datasets/silk-road/alpaca-data-gpt4-chinese |
|
|
- https://huggingface.co/datasets/LooksJuicy/ruozhiba |
|
|
|
|
|
|
|
|
### 结果 |
|
|
| 模型名称 | CEVAL | MMLU | |
|
|
|------------------------ |-------|------| |
|
|
| Qwen1.5-7B-Chat | 68.61 | 61.56| |
|
|
| Qwen1.5-7B-Chat-sft-lora-tigerbot-alpacadatagpt4-ruozhiba-1epoch | 71.75 | | |
|
|
|
|
|
|
|
|
|
