Instructions to use SUSTech/SUS-Chat-34B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SUSTech/SUS-Chat-34B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="SUSTech/SUS-Chat-34B")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("SUSTech/SUS-Chat-34B", dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use SUSTech/SUS-Chat-34B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "SUSTech/SUS-Chat-34B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "SUSTech/SUS-Chat-34B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/SUSTech/SUS-Chat-34B
- SGLang
How to use SUSTech/SUS-Chat-34B 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 "SUSTech/SUS-Chat-34B" \ --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": "SUSTech/SUS-Chat-34B", "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 "SUSTech/SUS-Chat-34B" \ --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": "SUSTech/SUS-Chat-34B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use SUSTech/SUS-Chat-34B with Docker Model Runner:
docker model run hf.co/SUSTech/SUS-Chat-34B
SusChat not generating the Output
Hi, I tried setting up the Model on my local computer. However, the Model does not perform exactly as it does on https://huggingface.co/spaces/SUSTech/SUS-Chat-34B. Sometimes it outputs the expected output but sometimes it does not output anything. Am I doing something incorrect with the code? Is the Model I downloaded the same as on the UI? This is the code I used:-
For context, I am opening the CSV file to read the value and compare it against the 1000 which is not currently implemented on the code for testing.
- Use prompt like ### Human: [your prompt here]\n\n###Assistant: for completion.
- HF online space version is run in bfloat16 precision instead of llama-cpp quantilized version, so there may exist some differences between the space and your local env (We run this model in VLLM).

Hi, Thank you for replying. I have used the code provided in the SusTech/ Sus-Chat's Model Card. When I run the same prompt in the UI, the accuracy of the output is 95% but when ran locally is 75%. I think the main problem is related to the chat template function. What should I need to make changes in this code to fix this issue?
Thanks.
