Update README.md

README.md

@@ -41,6 +41,7 @@ from vllm import LLM, SamplingParams
 from transformers import AutoTokenizer
 
 model_id = "neuralmagic/Meta-Llama-3-8B-Instruct-quantized.w8a16"
+number_gpus = 1
 
 sampling_params = SamplingParams(temperature=0.6, top_p=0.9, max_tokens=256)
 
@@ -53,7 +54,7 @@ messages = [
 
 prompts = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)
 
-llm = LLM(model=model_id, tensor_parallel_size=1)
+llm = LLM(model=model_id, tensor_parallel_size=number_gpus)
 
 outputs = llm.generate(prompts, sampling_params)