Instructions to use AnLan577/Dynamic_MoE with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AnLan577/Dynamic_MoE with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="AnLan577/Dynamic_MoE")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("AnLan577/Dynamic_MoE") model = AutoModelForCausalLM.from_pretrained("AnLan577/Dynamic_MoE") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use AnLan577/Dynamic_MoE with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "AnLan577/Dynamic_MoE" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "AnLan577/Dynamic_MoE", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/AnLan577/Dynamic_MoE
- SGLang
How to use AnLan577/Dynamic_MoE 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 "AnLan577/Dynamic_MoE" \ --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": "AnLan577/Dynamic_MoE", "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 "AnLan577/Dynamic_MoE" \ --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": "AnLan577/Dynamic_MoE", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use AnLan577/Dynamic_MoE with Docker Model Runner:
docker model run hf.co/AnLan577/Dynamic_MoE
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,18 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
---
|
| 4 |
+
Model weights For the Paper ""Harder Tasks Need More Experts: Dynamic Routing in MoE Models""
|
| 5 |
+
|
| 6 |
+
Inference Code can be found at: https://github.com/ZhenweiAn/Dynamic_MoE
|
| 7 |
+
|
| 8 |
+
@article{huang2024harder,
|
| 9 |
+
|
| 10 |
+
title={Harder Tasks Need More Experts: Dynamic Routing in MoE Models},
|
| 11 |
+
|
| 12 |
+
author={Huang, Quzhe and An, Zhenwei and Zhuang, Nan and Tao, Mingxu and Zhang, Chen and Jin, Yang and Xu, Kun and Chen, Liwei and Huang, Songfang and Feng, Yansong},
|
| 13 |
+
|
| 14 |
+
journal={arXiv preprint arXiv:2403.07652},
|
| 15 |
+
|
| 16 |
+
year={2024}
|
| 17 |
+
|
| 18 |
+
}
|