🤗 Hugging Face • 🤖 ModelScope • 👾 Wisemodel • 💬 WeChat• 📜Tech Report
# Demonstration of Hugging Face Model Inference
## Base Model Inference
We can perform inference for the Skywork-MoE-Base (16x13B size) model using HuggingFace on 8xA100/A800 or higher GPU hardware configurations.
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("Skywork/Skywork-MoE-Base", trust_remote_code=True, device_map='auto')
tokenizer = AutoTokenizer.from_pretrained("Skywork/Skywork-MoE-Base", trust_remote_code=True)
inputs = tokenizer('陕西的省会是西安', return_tensors='pt').to(model.device)
response = model.generate(inputs.input_ids, max_length=128)
print(tokenizer.decode(response.cpu()[0], skip_special_tokens=True))
"""
陕西的省会是西安。
西安，古称长安、镐京，是陕西省会、副省级市、关中平原城市群核心城市、丝绸之路起点城市、“一带一路”核心区、中国西部地区重要的中心城市，国家重要的科研、教育、工业基地。
西安是中国四大古都之一，联合国科教文组织于1981年确定的“世界历史名城”，美媒评选的世界十大古都之一。地处关中平原中部，北濒渭河，南依秦岭，八水润长安。下辖11区2县并代管西
"""
inputs = tokenizer('陕西的省会是西安，甘肃的省会是兰州，河南的省会是郑州', return_tensors='pt').to(model.device)
response = model.generate(inputs.input_ids, max_length=128)
print(tokenizer.decode(response.cpu()[0], skip_special_tokens=True))
"""
陕西的省会是西安，甘肃的省会是兰州，河南的省会是郑州，湖北的省会是武汉，湖南的省会是长沙，安徽的省会是合肥，江西的省会是南昌，江苏的省会是南京，浙江的省会是杭州，福建的省会是福州，广东的省会是广州，广西的省会是南宁，四川的省会是成都，贵州的省会是贵阳，云南的省会是昆明，山西的省会是太原，山东的省会是济南，河北的省会是石家庄，辽宁的省会是沈阳，吉林的省会是长春，黑龙江的
"""
```
## Chat Model Inference
coming soon...
# Demonstration of vLLM Model Inference
## Quickstart with vLLM
We provide a method to quickly deploy the Skywork-MoE-Base model based on vllm.
You can get the source code in [`vllm`](https://github.com/SkyworkAI/vllm)
### Based on local environment
Some dependencies need to be installed:
```shell
pip3 install xformers vllm-flash-attn
```
Then clone the [`vllm`](https://github.com/SkyworkAI/vllm) provided by skywork:
``` shell
git clone https://github.com/SkyworkAI/vllm.git
cd vllm
```
Then compile and install vllm:
``` shell
MAX_JOBS=8 python3 setup.py install
```
### Based on docker
You can use the docker image provided by skywork to run vllm directly:
```shell
docker pull registry.cn-wulanchabu.aliyuncs.com/triple-mu/skywork-moe-vllm:v1
```
Then start the container and set the model path and working directory.
```shell
model_path="Skywork/Skywork-MoE-Base"
workspace=${PWD}
docker run \
--runtime nvidia \
--gpus all \
-it \
--rm \
--shm-size=1t \
--ulimit memlock=-1 \
--privileged=true \
--ulimit stack=67108864 \
--ipc=host \
-v ${model_path}:/Skywork-MoE-Base \
-v ${workspace}:/workspace \
registry.cn-wulanchabu.aliyuncs.com/triple-mu/skywork-moe-vllm:v1
```
Now, you can run the Skywork MoE model for fun!
### Text Completion
``` python
from vllm import LLM, SamplingParams
model_path = 'Skywork/Skywork-MoE-Base'
prompts = [
"The president of the United States is",
"The capital of France is",
]
sampling_params = SamplingParams(temperature=0.3, max_tokens=256)
llm = LLM(
model=model_path,
kv_cache_dtype='auto',
tensor_parallel_size=8,
gpu_memory_utilization=0.95,
enforce_eager=True,
trust_remote_code=True,
)
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
```
# Declaration and License Agreement
## Declaration
We hereby declare that the Skywork model should not be used for any activities that pose a threat to national or societal security or engage in unlawful actions. Additionally, we request users not to deploy the Skywork model for internet services without appropriate security reviews and records. We hope that all users will adhere to this principle to ensure that technological advancements occur in a regulated and lawful environment.
We have done our utmost to ensure the compliance of the data used during the model's training process. However, despite our extensive efforts, due to the complexity of the model and data, there may still be unpredictable risks and issues. Therefore, if any problems arise as a result of using the Skywork open-source model, including but not limited to data security issues, public opinion risks, or any risks and problems arising from the model being misled, abused, disseminated, or improperly utilized, we will not assume any responsibility.
## License Agreement
The community usage of Skywork model requires [Skywork Community License](https://github.com/SkyworkAI/Skywork-MoE/blob/main/Skywork%20Community%20License.pdf). The Skywork model supports commercial use. If you plan to use the Skywork model or its derivatives for commercial purposes, you must abide by terms and conditions within [Skywork Community License](https://github.com/SkyworkAI/Skywork-MoE/blob/main/Skywork%20Community%20License.pdf).
[《Skywork 模型社区许可协议》》]:https://github.com/SkyworkAI/Skywork-MoE/blob/main/Skywork%20模型社区许可协议.pdf
[skywork-opensource@kunlun-inc.com]: mailto:skywork-opensource@kunlun-inc.com
# Contact Us and Citation
If you find our work helpful, please feel free to cite our paper~
```
@misc{wei2024skywork,
title={Skywork-MoE: A Deep Dive into Training Techniques for Mixture-of-Experts Language Models},
author={Tianwen Wei, Bo Zhu, Liang Zhao, Cheng Cheng, Biye Li, Weiwei Lü, Peng Cheng, Jianhao Zhang, Xiaoyu Zhang, Liang Zeng, Xiaokun Wang, Yutuan Ma, Rui Hu, Shuicheng Yan, Han Fang, Yahui Zhou},
year={2024},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
Paper: arxiv.org/abs/2406.06563