Instructions to use chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt")
model = AutoModelForCausalLM.from_pretrained("chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt

SGLang

How to use chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt with Docker Model Runner:
```
docker model run hf.co/chestnutlzj/MoE-Qwen-4x1.8B-pretrain-50000-ckpt
```

4x1.8B MoE Qwen Ckpt 50000

This is a MoE model project constructed based on the Qwen 1.8B model. In this project, we concatenated 4 original models and trained them using special training methods.

This model is a checkpoint model for the continue pretraining stage.

Evaluations

Groups	Metric	Value		Stderr
boolq	acc	0.6508	±	0.0083
ceval-valid	acc	0.5290	±	0.1912
	acc_norm	0.5290	±	0.1912
cmmlu	acc	0.5087	±	0.1237
	acc_norm	0.5087	±	0.1237
mathqa	acc	0.2647	±	0.0081
	acc_norm	0.2693	±	0.0081
mmlu	acc	0.4353	±	0.0830
- stem	acc	0.3809	±	0.0659
- social_sciences	acc	0.4959	±	0.0708
- other	acc	0.4844	±	0.0744
- humanities	acc	0.3998	±	0.0849

Acknowledgements

License Agreement

This project is open source under the Tongyi Qianwen Research License Agreement. You can view the complete license agreement in this link: [https://github.com/QwenLM/Qwen/blob/main/Tongyi%20Qianwen%20RESEARCH%20LICENSE%20AGREEMENT].

During the use of this project, please ensure that your usage behavior complies with the terms and conditions of the license agreement.

Downloads last month: 14

Safetensors

Model size

4B params

Tensor type

F32

BF16