diffusers/pr_14398/en/api/models/ace_step_transformer.md

AceStepTransformer1DModel

A 1D Diffusion Transformer for music generation from ACE-Step 1.5. The model operates on the 25 Hz stereo latents produced by AutoencoderOobleck using flow matching, and is trained with a Qwen3-derived backbone (grouped-query attention, rotary position embedding, RMSNorm, AdaLN-Zero timestep conditioning) plus cross-attention to the text / lyric / timbre conditions built by AceStepConditionEncoder.

AceStepTransformer1DModel[[diffusers.AceStepTransformer1DModel]]

diffusers.AceStepTransformer1DModel[[diffusers.AceStepTransformer1DModel]]

diffusers.AceStepTransformer1DModel(hidden_size: int = 2048, intermediate_size: int = 6144, num_hidden_layers: int = 24, num_attention_heads: int = 16, num_key_value_heads: int = 8, head_dim: int = 128, in_channels: int = 192, audio_acoustic_hidden_dim: int = 64, patch_size: int = 2, rope_theta: float = 1000000.0, attention_bias: bool = False, attention_dropout: float = 0.0, rms_norm_eps: float = 1e-06, sliding_window: int = 128, layer_types: typing.Optional[typing.List[str]] = None, encoder_hidden_size: typing.Optional[int] = None, is_turbo: bool = False, model_version: typing.Optional[str] = None)

Source

Diffusion Transformer for ACE-Step 1.5 music generation.

Generates audio latents conditioned on text, lyrics, and timbre. Uses 1D patch embedding (Conv1d with stride patch_size) followed by a stack of AceStepTransformerBlocks with alternating sliding-window / full attention on the self-attention branch. Cross-attention consumes the packed encoder_hidden_states produced by AceStepConditionEncoder.

forward[[diffusers.AceStepTransformer1DModel.forward]]

forward(hidden_states: Tensor, timestep: Tensor, timestep_r: Tensor, encoder_hidden_states: Tensor, context_latents: Tensor, attention_kwargs: typing.Optional[dict] = None, return_dict: bool = True)

Source

Parameters:

hidden_states (torch.Tensor of shape (batch_size, seq_len, channels)) : Noisy latent input for the diffusion process.

timestep (torch.Tensor of shape (batch_size,)) : Current diffusion timestep t.

timestep_r (torch.Tensor of shape (batch_size,)) : Reference timestep r (set equal to t for standard inference).

encoder_hidden_states (torch.Tensor of shape (batch_size, encoder_seq_len, hidden_size)) : Conditioning embeddings from the condition encoder (text + lyrics + timbre).

context_latents (torch.Tensor of shape (batch_size, seq_len, context_dim)) : Context latents (source latents concatenated with chunk masks) — fed to the patchify conv alongside hidden_states.

attention_kwargs (dict, optional) : A kwargs dictionary passed along to the AttentionProcessor. Used to pass the LoRA scale via {"scale": float}.

return_dict (bool, defaults to True) : Whether to return a Transformer2DModelOutput or a plain tuple.

Returns: Transformer2DModelOutput or tuple

The predicted velocity field.

The AceStepTransformer1DModel forward method.