88 lines
2.9 KiB
Markdown
88 lines
2.9 KiB
Markdown
|
---
|
|||
|
|
license: apache-2.0
|
|||
|
|
language:
|
|||
|
|
- en
|
|||
|
|
- zh
|
|||
|
|
pipeline_tag: text-generation
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
# Unichat-llama3-Chinese-8B-28K
|
|||
|
|
|
|||
|
|
|
|||
|
|
## 介绍
|
|||
|
|
* 中国联通AI创新中心发布业界第一个llama3中文指令微调模型长文本版本，全参数微调(非lora/longlora)，2024年4月26日19点
|
|||
|
|
* 本模型以[**Meta Llama 3**](https://huggingface.co/collections/meta-llama/meta-llama-3-66214712577ca38149ebb2b6)为基础,增加中文数据进行训练,实现llama3模型高质量中文问答
|
|||
|
|
* 此模型上下文原生长度28K
|
|||
|
|
* 基础模型 [**Meta-Llama-3-8B**](https://huggingface.co/meta-llama/Meta-Llama-3-8B)
|
|||
|
|
|
|||
|
|
|
|||
|
|
### 📊 数据
|
|||
|
|
- 高质量长文本指令数据，覆盖多个领域和行业，为模型训练提供充足的数据支持
|
|||
|
|
- 微调指令数据经过严格的人工筛查，保证优质的指令数据用于模型微调
|
|||
|
|
|
|||
|
|
## 基础示例
|
|||
|
|
|
|||
|
|
* 摘取《明朝那些事儿》小说第[319]章至[337]章内容约1万6千字左右，对文章内容进行提问。
|
|||
|
|
* 问:明朝三大营的组成及职责是什么？
|
|||
|
|
* 答:根据文章内容,明朝三大营的组成及职责如下:
|
|||
|
|
|
|||
|
|
1. 五军营:由骑兵和步兵组成,分为中军、左军、左掖军、右掖军、右哨军,担任攻击的主力。
|
|||
|
|
|
|||
|
|
2. 三千营:由投降的蒙古骑兵组成,主要是以雇佣兵为主的。下属全部都是骑兵,这支骑兵部队人数虽然不多,却是朱棣手下最为强悍的骑兵力量,他们在战争中主要担任突击的角色。
|
|||
|
|
|
|||
|
|
3. 神机营:使用火炮和火铳等火器为主的部队,是明朝政府的炮兵部队。朱棣同志之所以要组建这样的一支部队,那是有着深刻原因的。
|
|||
|
|
|
|||
|
|
所以,明朝三大营的组成和职责是:五军营负责攻击,三千营负责突击,神机营负责使用火器进行打击。
|
|||
|
|
|
|||
|
|
## 快速开始
|
|||
|
|
|
|||
|
|
```python
|
|||
|
|
import transformers
|
|||
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM
|
|||
|
|
import torch
|
|||
|
|
|
|||
|
|
model_id = "UnicomAI/Unichat-llama3-Chinese-8B-28K"
|
|||
|
|
|
|||
|
|
pipeline = transformers.pipeline(
|
|||
|
|
"text-generation",
|
|||
|
|
model=model_id,
|
|||
|
|
model_kwargs={"torch_dtype": torch.bfloat16},
|
|||
|
|
device="cuda",
|
|||
|
|
)
|
|||
|
|
|
|||
|
|
|
|||
|
|
messages = [
|
|||
|
|
{"role": "system", "content": "A chat between a curious user and an artificial intelligence assistant.The assistant gives helpful, detailed, and polite answers to the user's questions."},
|
|||
|
|
{"role": "user", "content": "你是谁"},
|
|||
|
|
]
|
|||
|
|
|
|||
|
|
|
|||
|
|
prompt = pipeline.tokenizer.apply_chat_template(
|
|||
|
|
messages,
|
|||
|
|
tokenize=False,
|
|||
|
|
add_generation_prompt=True
|
|||
|
|
)
|
|||
|
|
|
|||
|
|
terminators = [
|
|||
|
|
pipeline.tokenizer.eos_token_id,
|
|||
|
|
pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
|
|||
|
|
]
|
|||
|
|
|
|||
|
|
|
|||
|
|
outputs = pipeline(
|
|||
|
|
prompt,
|
|||
|
|
max_new_tokens=2048,
|
|||
|
|
eos_token_id=terminators,
|
|||
|
|
do_sample=True,
|
|||
|
|
temperature=0.6,
|
|||
|
|
top_p=1,
|
|||
|
|
repetition_penalty=1.15
|
|||
|
|
)
|
|||
|
|
print(outputs[0]["generated_text"][len(prompt):])
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
## 资源
|
|||
|
|
更多模型,数据集和训练相关细节请参考:
|
|||
|
|
* Github：[**Unichat-llama3-Chinese**](https://github.com/UnicomAI/Unichat-llama3-Chinese)
|
|||
|
|
|