> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/automodel/v0.5/nemo-automodel/nemo_automodel/components/models/ling_v2/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.ling_v2

## Submodules

* **[`nemo_automodel.components.models.ling_v2.config`](/nemo-automodel/nemo_automodel/components/models/ling_v2/config)**
* **[`nemo_automodel.components.models.ling_v2.layers`](/nemo-automodel/nemo_automodel/components/models/ling_v2/layers)**
* **[`nemo_automodel.components.models.ling_v2.model`](/nemo-automodel/nemo_automodel/components/models/ling_v2/model)**
* **[`nemo_automodel.components.models.ling_v2.state_dict_adapter`](/nemo-automodel/nemo_automodel/components/models/ling_v2/state_dict_adapter)**

## Package Contents

### Classes

| Name                                                                                                 | Description                                                |
| ---------------------------------------------------------------------------------------------------- | ---------------------------------------------------------- |
| [`BailingMoeV2Config`](#nemo_automodel-components-models-ling_v2-config-BailingMoeV2Config)          | Configuration class for the BailingMoeV2 model (Ling 2.0). |
| [`BailingMoeV2ForCausalLM`](#nemo_automodel-components-models-ling_v2-model-BailingMoeV2ForCausalLM) | Causal-LM head wrapping `BailingMoeV2Model`.               |

### API

```python
class nemo_automodel.components.models.ling_v2.config.BailingMoeV2Config(
    vocab_size: int = 157184,
    hidden_size: int = 2048,
    intermediate_size: int = 5120,
    num_hidden_layers: int = 20,
    num_attention_heads: int = 16,
    num_key_value_heads: int = 4,
    hidden_act: str = 'silu',
    use_qkv_bias: bool = False,
    use_bias: bool = False,
    rms_norm_eps: float = 1e-06,
    tie_word_embeddings: bool = False,
    embedding_dropout: float = 0.0,
    attention_dropout: float = 0.0,
    output_dropout: float = 0.0,
    initializer_range: float = 0.02,
    max_position_embeddings: int = 32768,
    rope_theta: float = 600000.0,
    use_cache: bool = True,
    max_window_layers: int = 20,
    rope_scaling: dict | None = None,
    pad_token_id: int = 156892,
    eos_token_id: int = 156892,
    num_experts: int = 256,
    num_shared_experts: int = 1,
    num_experts_per_tok: int = 8,
    n_group: int = 8,
    topk_group: int = 4,
    moe_intermediate_size: int = 512,
    first_k_dense_replace: int = 1,
    head_dim: int = 128,
    output_router_logits: bool = False,
    use_qk_norm: bool = True,
    partial_rotary_factor: float = 1.0,
    num_nextn_predict_layers: int = 0,
    mtp_loss_scaling_factor: float = 0,
    moe_router_enable_expert_bias: bool = True,
    routed_scaling_factor: float = 1.0,
    norm_topk_prob: bool = True,
    score_function: str = 'sigmoid',
    rotary_dim: int | None = None,
    kwargs = {}
)
```

**Bases:** `PretrainedConfig`

Configuration class for the BailingMoeV2 model (Ling 2.0).

The defaults reflect the `Ling-mini-2.0` (16B-A1.4B) variant.  Larger
variants (`Ling-flash-2.0` 100B-A6B and `Ling-1T` 1T-A50B) override
sizing knobs but share the same architecture: GQA attention with per-head
QK-RMSNorm, partial RoPE, sigmoid-routed grouped MoE with shared experts,
and `first_k_dense_replace` dense MLP layers at the start.

**`head_dim`** `= head_dim or hidden_size // num_attention_heads`

---

**`keys_to_ignore_at_inference`** `= ['past_key_values']`

---

**`model_type`** `= 'bailing_moe'`

---

```python
class nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM(
    config: nemo_automodel.components.models.ling_v2.config.BailingMoeV2Config,
    moe_config: nemo_automodel.components.moe.config.MoEConfig | None = None,
    backend: nemo_automodel.components.models.common.BackendConfig | None = None,
    kwargs = {}
)
```

**Bases:** [HFCheckpointingMixin](/nemo-automodel/nemo_automodel/components/models/common/hf_checkpointing_mixin#nemo_automodel-components-models-common-hf_checkpointing_mixin-HFCheckpointingMixin), `Module`, [MoEFSDPSyncMixin](/nemo-automodel/nemo_automodel/components/moe/fsdp_mixin#nemo_automodel-components-moe-fsdp_mixin-MoEFSDPSyncMixin)

Causal-LM head wrapping `BailingMoeV2Model`.

**`_keep_in_fp32_modules_strict`** `= ['e_score_correction_bias', 'expert_bias']`

---

**`_pp_keep_self_forward`** `bool = True`

---

**`lm_head`**

---

**`model`**

---

**`state_dict_adapter`**

---

**`tie_word_embeddings_support`** `TieSupport = TieSupport.UNTIED_ONLY`

---

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.forward(
    input_ids: torch.Tensor,
    position_ids: torch.Tensor | None = None,
    attention_mask: torch.Tensor | None = None,
    padding_mask: torch.Tensor | None = None,
    logits_to_keep: typing.Union[int, torch.Tensor] = 0,
    output_hidden_states: bool | None = None,
    attn_kwargs: typing.Any = {}
) -> transformers.modeling_outputs.CausalLMOutputWithPast
```

Forward pass returning `CausalLMOutputWithPast`.

Supports BSHD (`input_ids` shape `[B, S]`) and THD (squeezed to `[T]`
when `attn_kwargs["qkv_format"] == "thd"`) formats.

**Parameters:**

**`input_ids`** `torch.Tensor`

Input token IDs.

---

**`position_ids`** `torch.Tensor | None` — default: None

Optional position indices.

---

**`attention_mask`** `torch.Tensor | None` — default: None

Optional 2D padding mask.

---

**`padding_mask`** `torch.Tensor | None` — default: None

Optional padding mask used by the THD squeeze helper.

---

**`logits_to_keep`** `Union[int, torch.Tensor]` — default: 0

If > 0, only compute logits for the last
`logits_to_keep` positions (avoids materialising the full logit
matrix during generation / fused-CE training). `0` computes all
positions.

---

**`output_hidden_states`** `bool | None` — default: None

Whether to return the final hidden states (the
input to `lm_head`) on the output. Required by the fused
cross-entropy (cut-CE) training path.

---

**`**attn_kwargs`** `Any` — default: \{}

Additional arguments forwarded to the base model.

---

**Returns:** `CausalLMOutputWithPast`

class:`~transformers.modeling_outputs.CausalLMOutputWithPast` with

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.from_config(
    config: nemo_automodel.components.models.ling_v2.config.BailingMoeV2Config,
    moe_config: nemo_automodel.components.moe.config.MoEConfig | None = None,
    backend: nemo_automodel.components.models.common.BackendConfig | None = None,
    kwargs = {}
)
```

classmethod

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.from_pretrained(
    pretrained_model_name_or_path: str,
    model_args = (),
    kwargs = {}
)
```

classmethod

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.get_capabilities(
    config
) -> nemo_automodel._transformers.model_capabilities.ModelCapabilities
```

classmethod

Return parallelism capabilities for a specific Ling/Bailing-MoE config.

Three checkpoint variants share this class:

1. `inclusionAI/Ling-1T` -- 1T-param MoE, requires PP.
   Demonstrated by examples/llm\_finetune/ling/ling\_1t\_sft.yaml (pp=4)
   and ling\_1t\_lora\_pp.yaml (pp=8); both with ep\_size>=8.
2. `inclusionAI/Ling-flash-2.0` -- mid-size MoE, single-rank EP only.
   Demonstrated by ling\_flash\_2\_0\_sft.yaml / ling\_flash\_2\_0\_lora.yaml
   (pp=1, ep=8-32).
3. `inclusionAI/Ling-mini-2.0` -- small MoE, single-rank EP only.
   Demonstrated by ling\_mini\_2\_0\_\{hellaswag,sft,squad}.yaml
   (pp=1, ep=4-8).

Dispatch is on num\_hidden\_layers since Ling-1T (\~80 layers) is well
separated from Ling-flash-2.0 (\~32) and Ling-mini-2.0 (\~20).

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.get_input_embeddings()
```

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.get_output_embeddings()
```

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.initialize_weights(
    buffer_device: torch.device | None = None,
    dtype: torch.dtype = torch.bfloat16
) -> None
```

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.set_input_embeddings(
    value
)
```

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.set_output_embeddings(
    new_embeddings
)
```

```python
nemo_automodel.components.models.ling_v2.model.BailingMoeV2ForCausalLM.update_moe_gate_bias() -> None
```

## Docs

- [nemo_automodel.components.models.ling_v2.config](https://docs.nvidia.com/nemo/automodel/v0.5/nemo-automodel/nemo_automodel/components/models/ling_v2/config.md)
- [nemo_automodel.components.models.ling_v2.layers](https://docs.nvidia.com/nemo/automodel/v0.5/nemo-automodel/nemo_automodel/components/models/ling_v2/layers.md)
- [nemo_automodel.components.models.ling_v2.model](https://docs.nvidia.com/nemo/automodel/v0.5/nemo-automodel/nemo_automodel/components/models/ling_v2/model.md)
- [nemo_automodel.components.models.ling_v2.state_dict_adapter](https://docs.nvidia.com/nemo/automodel/v0.5/nemo-automodel/nemo_automodel/components/models/ling_v2/state_dict_adapter.md)