***

layout: overview
slug: nemo-curator/nemo\_curator/models/vllm\_model
title: nemo\_curator.models.vllm\_model
---------------------------------------

## Module Contents

### Classes

| Name                                                               | Description                                              |
| ------------------------------------------------------------------ | -------------------------------------------------------- |
| [`LLM`](#nemo_curator-models-vllm_model-LLM)                       | -                                                        |
| [`SamplingParams`](#nemo_curator-models-vllm_model-SamplingParams) | -                                                        |
| [`VLLMModel`](#nemo_curator-models-vllm_model-VLLMModel)           | Generic vLLM language model wrapper for text generation. |

### Data

[`VLLM_AVAILABLE`](#nemo_curator-models-vllm_model-VLLM_AVAILABLE)

### API

<Anchor id="nemo_curator-models-vllm_model-LLM">
  <CodeBlock showLineNumbers={false} wordWrap={true}>
    ```python
    class nemo_curator.models.vllm_model.LLM()
    ```
  </CodeBlock>
</Anchor>

<Indent />

<Anchor id="nemo_curator-models-vllm_model-SamplingParams">
  <CodeBlock showLineNumbers={false} wordWrap={true}>
    ```python
    class nemo_curator.models.vllm_model.SamplingParams()
    ```
  </CodeBlock>
</Anchor>

<Indent />

<Anchor id="nemo_curator-models-vllm_model-VLLMModel">
  <CodeBlock showLineNumbers={false} wordWrap={true}>
    ```python
    class nemo_curator.models.vllm_model.VLLMModel(
        model: str,
        max_model_len: int | None = None,
        tensor_parallel_size: int | None = None,
        max_num_batched_tokens: int = 4096,
        temperature: float = 0.7,
        top_p: float = 0.8,
        top_k: int = 20,
        min_p: float = 0.0,
        max_tokens: int | None = None,
        cache_dir: str | None = None
    )
    ```
  </CodeBlock>
</Anchor>

<Indent>
  **Bases:** [ModelInterface](/nemo-curator/nemo_curator/models/base#nemo_curator-models-base-ModelInterface)

  Generic vLLM language model wrapper for text generation.

  <ParamField path="_final_max_model_len" type="int | None = None" />

  <ParamField path="_is_qwen3" type="bool = False" />

  <ParamField path="_llm" type="LLM | None = None" />

  <ParamField path="_sampling_params" type="SamplingParams | None = None" />

  <ParamField path="model_id_names" type="list[str]">
    Return the model identifier.
  </ParamField>

  <Anchor id="nemo_curator-models-vllm_model-VLLMModel-generate">
    <CodeBlock showLineNumbers={false} wordWrap={true}>
      ```python
      nemo_curator.models.vllm_model.VLLMModel.generate(
          prompts: list[str]
      ) -> list[str]
      ```
    </CodeBlock>
  </Anchor>

  <Indent>
    Generate text from prompts.

    **Parameters:**

    <ParamField path="prompts" type="list[str]">
      List of prompt strings or list of message dicts
      (for chat template).
    </ParamField>

    **Returns:** `list[str]`

    List of generated text strings.

    **Raises:**

    * `RuntimeError`: If the model is not set up or generation fails.
  </Indent>

  <Anchor id="nemo_curator-models-vllm_model-VLLMModel-get_tokenizer">
    <CodeBlock showLineNumbers={false} wordWrap={true}>
      ```python
      nemo_curator.models.vllm_model.VLLMModel.get_tokenizer() -> typing.Any
      ```
    </CodeBlock>
  </Anchor>

  <Indent>
    Get the tokenizer from the LLM instance.
  </Indent>

  <Anchor id="nemo_curator-models-vllm_model-VLLMModel-setup">
    <CodeBlock showLineNumbers={false} wordWrap={true}>
      ```python
      nemo_curator.models.vllm_model.VLLMModel.setup() -> None
      ```
    </CodeBlock>
  </Anchor>

  <Indent>
    Set up the vLLM model and sampling parameters.
  </Indent>
</Indent>

<Anchor id="nemo_curator-models-vllm_model-VLLM_AVAILABLE">
  <CodeBlock showLineNumbers={false} wordWrap={true}>
    ```python
    nemo_curator.models.vllm_model.VLLM_AVAILABLE = True
    ```
  </CodeBlock>
</Anchor>