nemo_automodel.components.models.qwen3_next.layers

Module Contents

Classes

Name	Description
`Qwen3NextAttention`	-
`Qwen3NextFp32GatedDeltaNet`	Qwen3-Next GatedDeltaNet that computes the decay gate via an fp32 holder.
`Qwen3NextRMSNorm`	-
`Qwen3NextSSMGate`	Owns Qwen3-Next fp32 SSM-gating params and computes the decay gate.
`_SSMGateParam`	Get-only descriptor exposing a param from `_fp32_params` when present.

Functions

Name	Description
`_install_ssm_gate`	Move HF-created bare `A_log`/`dt_bias` into a native fp32 holder.

API

class nemo_automodel.components.models.qwen3_next.layers.Qwen3NextAttention(
    config: transformers.models.qwen3_next.configuration_qwen3_next.Qwen3NextConfig,
    layer_idx: int,
    backend: nemo_automodel.components.models.common.BackendConfig
)

Bases: Module

head_dim

k_norm

k_proj

num_heads

= config.num_attention_heads

num_key_value_groups

= self.num_heads // self.num_kv_heads

num_kv_heads

= config.num_key_value_heads

o_proj

q_norm

q_proj

scaling

= self.head_dim ** -0.5

v_proj

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextAttention.forward(
    x: torch.Tensor,
    freqs_cis: torch.Tensor,
    attention_mask: torch.Tensor | None = None,
    attn_kwargs: typing.Any = {}
) -> torch.Tensor

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextAttention.init_weights(
    buffer_device: torch.device,
    init_std: float = 0.02
)

class nemo_automodel.components.models.qwen3_next.layers.Qwen3NextFp32GatedDeltaNet(
    config: transformers.models.qwen3_next.configuration_qwen3_next.Qwen3NextConfig,
    layer_idx: int
)

Bases: Qwen3NextGatedDeltaNet

Qwen3-Next GatedDeltaNet that computes the decay gate via an fp32 holder.

HF’s Qwen3NextGatedDeltaNet computes the gate inline as g = -exp(A_log) * softplus(a + dt_bias) using the bare A_log / dt_bias parameters. A_log and dt_bias are intrinsically fp32 (A_log is exponentiated, so bf16 rounding becomes a proportional error on the decay rate that the recurrence compounds across the sequence).

The constructor moves those params into a native _fp32_params holder so they are fp32 resident before any dtype cast or FSDP wrapping. To keep the gate computation in fp32 — and to make FSDP’s unshard/reshard + gradient reduce-scatter fire for that unit — the gate is computed inside the holder’s forward. This subclass overrides forward to route the gate through self._compute_gate(a) while reproducing the rest of HF’s forward verbatim.

A_log

= _SSMGateParam('A_log')

dt_bias

= _SSMGateParam('dt_bias')

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextFp32GatedDeltaNet._bind_kernels() -> None

Restore the kernel callables HF dropped as instance attributes in 5.15.

transformers <=5.12 set causal_conv1d_fn / causal_conv1d_update / chunk_gated_delta_rule / recurrent_gated_delta_rule on the module and left causal_conv1d_fn as None when the kernel package was absent. 5.15 moved them to decorated module-level fallbacks, so forward below — which calls them with the kernel package’s own signature (x=, seq_idx=) — raised AttributeError. Rebind the installed kernels, and keep None when causal_conv1d is missing so the pure-torch conv branch in forward is taken rather than a signature-incompatible fallback.

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextFp32GatedDeltaNet._compute_gate(
    a: torch.Tensor
) -> torch.Tensor

Compute the decay gate g in fp32, via the holder when it exists.

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextFp32GatedDeltaNet.forward(
    hidden_states: torch.Tensor,
    cache_params: typing.Any | None = None,
    attention_mask: torch.Tensor | None = None
)

class nemo_automodel.components.models.qwen3_next.layers.Qwen3NextRMSNorm(
    dim: int,
    eps: float = 1e-06
)

Bases: Module

weight

= nn.Parameter(torch.zeros(dim))

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextRMSNorm._norm(
    x
)

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextRMSNorm.extra_repr()

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextRMSNorm.forward(
    x
)

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextRMSNorm.reset_parameters()

class nemo_automodel.components.models.qwen3_next.layers.Qwen3NextSSMGate(
    num_v_heads: int,
    dtype: torch.dtype = torch.float32
)

Bases: Module

Owns Qwen3-Next fp32 SSM-gating params and computes the decay gate.

A_log

dt_bias

nemo_automodel.components.models.qwen3_next.layers.Qwen3NextSSMGate.forward(
    a: torch.Tensor
) -> torch.Tensor

class nemo_automodel.components.models.qwen3_next.layers._SSMGateParam(
    name: str
)

Get-only descriptor exposing a param from _fp32_params when present.

nemo_automodel.components.models.qwen3_next.layers._SSMGateParam.__get__(
    obj,
    owner = None
)

nemo_automodel.components.models.qwen3_next.layers._install_ssm_gate(
    mod: torch.nn.Module,
    fp32_dtype: torch.dtype = torch.float32
) -> nemo_automodel.components.models.qwen3_next.layers.Qwen3NextSSMGate

Move HF-created bare A_log/dt_bias into a native fp32 holder.