> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/automodel/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/automodel/_mcp/server.

# nemo_automodel.components.datasets.vlm.mock

Mock VLM conversation dataset for benchmarking and testing.

Generates synthetic image(s) and minimal conversations in the standard
Automodel conversation format, compatible with `PreTokenizedDatasetWrapper`
and any HF `AutoProcessor` that supports the conversation schema.

The images are random-noise PIL images — no real data download is needed.
The processor / vision encoder processes them through the normal pipeline,
so this exercises the full VLM training path end-to-end.

When used with `pretokenize: true`, `truncate: true`, and `max_length`
in the dataset config, `PreTokenizedDatasetWrapper` tokenizes each sample
and truncates to exactly `max_length` tokens.  The mock response is
sized from `max_length` so that truncation always produces a full-length
sequence.

## Module Contents

### Classes

| Name                                                                                        | Description                                               |
| ------------------------------------------------------------------------------------------- | --------------------------------------------------------- |
| [`MockVlmDatasetConfig`](#nemo_automodel-components-datasets-vlm-mock-MockVlmDatasetConfig) | Construction-time configuration for the mock VLM dataset. |

### Functions

| Name                                                                                            | Description                                                        |
| ----------------------------------------------------------------------------------------------- | ------------------------------------------------------------------ |
| [`_generate_response`](#nemo_automodel-components-datasets-vlm-mock-_generate_response)         | Generate a dummy response of *num\_words* words from a fixed pool. |
| [`_make_random_image`](#nemo_automodel-components-datasets-vlm-mock-_make_random_image)         | Create a random-noise RGB PIL image.                               |
| [`build_mock_vlm_dataset`](#nemo_automodel-components-datasets-vlm-mock-build_mock_vlm_dataset) | Build a mock VLM dataset in Automodel conversation format.         |

### Data

[`_WORD_POOL`](#nemo_automodel-components-datasets-vlm-mock-_WORD_POOL)

### API

```python
class nemo_automodel.components.datasets.vlm.mock.MockVlmDatasetConfig(
    num_samples: int = 10,
    num_images_per_sample: int = 1,
    image_size: tuple[int, int] = (256, 256),
    prompt: str = 'Describe this image.',
    responses: list[str] | None = None,
    max_length: int | None = None,
    seed: int = 0
)
```

Dataclass

Construction-time configuration for the mock VLM dataset.

`(width, height)` of each generated image.

Target sequence length. When set (and `responses` is `None`), drives response word count.

Number of random noise images per user turn.

Number of synthetic conversation examples to generate.

Text prompt appended after the image(s) in the user turn.

Optional list of assistant responses, cycled over samples.

Random seed for reproducibility.

```python
nemo_automodel.components.datasets.vlm.mock.MockVlmDatasetConfig.build() -> list[dict[str, object]]
```

Build the mock VLM dataset from this config.

```python
nemo_automodel.components.datasets.vlm.mock._generate_response(
    rng: numpy.random.Generator,
    num_words: int
) -> str
```

Generate a dummy response of *num\_words* words from a fixed pool.

```python
nemo_automodel.components.datasets.vlm.mock._make_random_image(
    rng: numpy.random.Generator,
    size: typing.Tuple[int, int] = (256, 256)
) -> PIL.Image.Image
```

Create a random-noise RGB PIL image.

```python
nemo_automodel.components.datasets.vlm.mock.build_mock_vlm_dataset(
    num_samples: int = 10,
    num_images_per_sample: int = 1,
    image_size: typing.Tuple[int, int] = (256, 256),
    prompt: str = 'Describe this image.',
    responses: typing.Optional[typing.List[str]] = None,
    max_length: typing.Optional[int] = None,
    seed: int = 0,
    kwargs = {}
) -> list
```

Build a mock VLM dataset in Automodel conversation format.

Each sample is a dict with a `"conversation"` key whose value is a list
of user/assistant message dicts.  User messages contain one or more
`{"type": "image", "image": <PIL.Image>}` items followed by a text prompt.
Assistant messages contain a single text response.

This is the same format produced by `make_rdr_dataset`,
`make_unimm_chat_dataset`, and `make_meta_dataset`, so the returned
list can be fed directly to `PreTokenizedDatasetWrapper`.

When `max_length` is set and `responses` is `None`, each sample's
assistant response is generated with `max_length` words — guaranteed
to exceed `max_length` tokens so that `PreTokenizedDatasetWrapper`
with `truncate=True` produces exactly `max_length` tokens per sample.

**Parameters:**

Number of conversation examples to generate.

Number of random images per user turn.

`(width, height)` of each generated image.

Text prompt appended after the image(s) in the user turn.

Optional list of assistant responses.  Cycled over samples.

Target sequence length.  When set (and `responses` is
`None`), generates a response of `max_length` words per sample
so the tokenized sequence always exceeds `max_length` tokens.

Random seed for reproducibility.

**Returns:** `list`

A list of dicts, each with a single `"conversation"` key.

```python
nemo_automodel.components.datasets.vlm.mock._WORD_POOL = 'the image shows a landscape with mountains and rivers flowing through green val...
```