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

# nemo_curator.stages.audio.inference.asr.asr_nemo

## Module Contents

### Classes

| Name                                                                                               | Description                                          |
| -------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
| [`InferenceAsrNemoStage`](#nemo_curator-stages-audio-inference-asr-asr_nemo-InferenceAsrNemoStage) | Speech recognition inference using a NeMo ASR model. |

### API

```python
class nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage(
    name: str = 'ASR_inference',
    model_name: str = '',
    cache_dir: str | None = None,
    asr_model: typing.Any | None = None,
    filepath_key: str = 'audio_filepath',
    pred_text_key: str = 'pred_text',
    resources: nemo_curator.stages.resources.Resources = (lambda: Resources(cpus=1.0...,
    batch_size: int = 16
)
```

Dataclass

**Bases:** [ProcessingStage\[AudioTask, AudioTask\]](/nemo-curator/nemo_curator/stages/base#nemo_curator-stages-base-ProcessingStage)

Speech recognition inference using a NeMo ASR model.

Overrides `process_batch` for batched GPU inference.

**Parameters:**

Pretrained NeMo ASR model name.
See full list at [https://docs.nvidia.com/nemo-framework/user-guide/latest/nemotoolkit/asr/all\_chkpt.html](https://docs.nvidia.com/nemo-framework/user-guide/latest/nemotoolkit/asr/all_chkpt.html)

Optional directory for model download cache.
When set, NeMo stores/loads the pretrained checkpoint here
instead of the default cache location.

Key in the entry dict pointing to the audio file.

Key where the predicted transcription is stored.

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.__post_init__() -> None
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.check_cuda() -> torch.device
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.inputs() -> tuple[list[str], list[str]]
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.outputs() -> tuple[list[str], list[str]]
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.process(
    task: nemo_curator.tasks.AudioTask
) -> nemo_curator.tasks.AudioTask
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.process_batch(
    tasks: list[nemo_curator.tasks.AudioTask]
) -> list[nemo_curator.tasks.AudioTask]
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.setup(
    _worker_metadata: nemo_curator.backends.base.WorkerMetadata | None = None
) -> None
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.setup_on_node(
    _node_info: nemo_curator.backends.base.NodeInfo | None = None,
    _worker_metadata: nemo_curator.backends.base.WorkerMetadata | None = None
) -> None
```

```python
nemo_curator.stages.audio.inference.asr.asr_nemo.InferenceAsrNemoStage.transcribe(
    files: list[str]
) -> list[str]
```