diffusers/pr_12229/en/api/models/latte_transformer3d.md

LatteTransformer3DModel

A Diffusion Transformer model for 3D data from Latte.

LatteTransformer3DModel[[diffusers.LatteTransformer3DModel]]

class diffusers.LatteTransformer3DModeldiffusers.LatteTransformer3DModelhttps://github.com/huggingface/diffusers/blob/vr_12229/src/diffusers/models/transformers/latte_transformer_3d.py#L370[{"name": "num_attention_heads", "val": ": int = 16"}, {"name": "attention_head_dim", "val": ": int = 88"}, {"name": "in_channels", "val": ": typing.Optional[int] = None"}, {"name": "out_channels", "val": ": typing.Optional[int] = None"}, {"name": "num_layers", "val": ": int = 1"}, {"name": "dropout", "val": ": float = 0.0"}, {"name": "cross_attention_dim", "val": ": typing.Optional[int] = None"}, {"name": "attention_bias", "val": ": bool = False"}, {"name": "sample_size", "val": ": int = 64"}, {"name": "patch_size", "val": ": typing.Optional[int] = None"}, {"name": "activation_fn", "val": ": str = 'geglu'"}, {"name": "num_embeds_ada_norm", "val": ": typing.Optional[int] = None"}, {"name": "norm_type", "val": ": str = 'layer_norm'"}, {"name": "norm_elementwise_affine", "val": ": bool = True"}, {"name": "norm_eps", "val": ": float = 1e-05"}, {"name": "caption_channels", "val": ": int = None"}, {"name": "video_length", "val": ": int = 16"}]

forwarddiffusers.LatteTransformer3DModel.forwardhttps://github.com/huggingface/diffusers/blob/vr_12229/src/diffusers/models/transformers/latte_transformer_3d.py#L509[{"name": "hidden_states", "val": ": Tensor"}, {"name": "timestep", "val": ": typing.Optional[torch.LongTensor] = None"}, {"name": "encoder_hidden_states", "val": ": typing.Optional[torch.Tensor] = None"}, {"name": "encoder_attention_mask", "val": ": typing.Optional[torch.Tensor] = None"}, {"name": "enable_temporal_attentions", "val": ": bool = True"}, {"name": "return_dict", "val": ": bool = True"}]- hidden_states shape (batch size, channel, num_frame, height, width) -- Input hidden_states.

• timestep ( torch.LongTensor, optional) -- Used to indicate denoising step. Optional timestep to be applied as an embedding in AdaLayerNorm.

• encoder_hidden_states ( torch.FloatTensor of shape (batch size, sequence len, embed dims), optional) -- Conditional embeddings for cross attention layer. If not given, cross-attention defaults to self-attention.

• encoder_attention_mask ( torch.Tensor, optional) -- Cross-attention mask applied to encoder_hidden_states. Two formats supported:

  • Mask (batcheight, sequence_length) True = keep, False = discard.
  • Bias (batcheight, 1, sequence_length) 0 = keep, -10000 = discard.

  If ndim == 2: will be interpreted as a mask, then converted into a bias consistent with the format above. This bias will be added to the cross-attention scores.

• enable_temporal_attentions -- (bool, optional, defaults to True): Whether to enable temporal attentions.

• return_dict (bool, optional, defaults to True) -- Whether or not to return a ~models.unet_2d_condition.UNet2DConditionOutput instead of a plain tuple.0If return_dict is True, an ~models.transformer_2d.Transformer2DModelOutput is returned, otherwise a tuple where the first element is the sample tensor.

The LatteTransformer3DModel forward method.