Аренда GPU для vLLM
запуск за 5 минут,
от 168 ₽/час

python3 -m venv /opt/vllm && source /opt/vllm/bin/activate
pip install --upgrade 'vllm>=0.20.0'
vllm serve Qwen/Qwen3-32B \
  --tensor-parallel-size 1 \
  --max-model-len 32768 \
  --gpu-memory-utilization 0.90 \
  --enable-auto-tool-choice \
  --tool-call-parser hermes \
  --reasoning-parser qwen3 \
  --host 0.0.0.0 --port 8000

# NVIDIA Container Toolkit должен быть установлен
docker run --runtime nvidia --gpus all \
  -v ~/.cache/huggingface:/root/.cache/huggingface \
  --env HF_TOKEN=$HF_TOKEN \
  -p 8000:8000 --ipc=host \
  vllm/vllm-openai:latest \
  --model meta-llama/Llama-3.3-70B-Instruct \
  --tensor-parallel-size 2 \
  --quantization fp8 \
  --max-model-len 16384

# Python через openai SDK
pip install openai
python3 - <<'PY'
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:8000/v1", api_key="EMPTY")
resp = client.chat.completions.create(
    model="Qwen/Qwen3-32B",
    messages=[{"role": "user", "content": "Привет!"}],
)
print(resp.choices[0].message.content)
PY

# Или curl:
curl http://127.0.0.1:8000/v1/chat/completions \
  -H 'Content-Type: application/json' \
  -d '{"model":"Qwen/Qwen3-32B","messages":[{"role":"user","content":"hi"}]}'

python3 - <<'PY'
from vllm import LLM, SamplingParams
llm = LLM(model="Qwen/Qwen3-8B", gpu_memory_utilization=0.90)
params = SamplingParams(temperature=0.7, max_tokens=512)
prompts = ["Объясни PagedAttention", "Что такое continuous batching?"]
outputs = llm.generate(prompts, params)
for o in outputs:
    print(o.outputs[0].text)
PY