> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/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.models.gemma4_moe

## Submodules

* **[`nemo_automodel.components.models.gemma4_moe.cp_attention`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/cp_attention)**
* **[`nemo_automodel.components.models.gemma4_moe.cp_batch`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/cp_batch)**
* **[`nemo_automodel.components.models.gemma4_moe.cp_local_ring`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/cp_local_ring)**
* **[`nemo_automodel.components.models.gemma4_moe.loss`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/loss)**
* **[`nemo_automodel.components.models.gemma4_moe.model`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/model)**
* **[`nemo_automodel.components.models.gemma4_moe.parallelization`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/parallelization)**
* **[`nemo_automodel.components.models.gemma4_moe.sdpa_fp32`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/sdpa_fp32)**
* **[`nemo_automodel.components.models.gemma4_moe.state_dict_adapter`](/nemo-automodel/nemo_automodel/components/models/gemma4_moe/state_dict_adapter)**

## Docs

- [nemo_automodel.components.models.gemma4_moe.cp_attention](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/cp_attention.md)
- [nemo_automodel.components.models.gemma4_moe.cp_batch](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/cp_batch.md)
- [nemo_automodel.components.models.gemma4_moe.cp_local_ring](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/cp_local_ring.md)
- [nemo_automodel.components.models.gemma4_moe.loss](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/loss.md)
- [nemo_automodel.components.models.gemma4_moe.model](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/model.md)
- [nemo_automodel.components.models.gemma4_moe.parallelization](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/parallelization.md)
- [nemo_automodel.components.models.gemma4_moe.sdpa_fp32](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/sdpa_fp32.md)
- [nemo_automodel.components.models.gemma4_moe.state_dict_adapter](https://docs.nvidia.com/nemo/automodel/v0.4/nemo-automodel/nemo_automodel/components/models/gemma4_moe/state_dict_adapter.md)