How to use SparseLLM/swiglu-40B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="SparseLLM/swiglu-40B")
# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("SparseLLM/swiglu-40B") model = AutoModelForCausalLM.from_pretrained("SparseLLM/swiglu-40B", device_map="auto")
How to use SparseLLM/swiglu-40B with vLLM:
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "SparseLLM/swiglu-40B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "SparseLLM/swiglu-40B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'
docker model run hf.co/SparseLLM/swiglu-40B
How to use SparseLLM/swiglu-40B with SGLang:
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "SparseLLM/swiglu-40B" \ --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": "SparseLLM/swiglu-40B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'
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 "SparseLLM/swiglu-40B" \ --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": "SparseLLM/swiglu-40B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'
How to use SparseLLM/swiglu-40B with Docker Model Runner:
