Transformer2D
Last updated
Last updated
A Transformer model for image-like data from that is based on the introduced by Dosovitskiy et al. The accepts discrete (classes of vector embeddings) or continuous (actual embeddings) inputs.
When the input is continuous:
Project the input and reshape it to (batch_size, sequence_length, feature_dimension).
Apply the Transformer blocks in the standard way.
Reshape to image.
When the input is discrete:
It is assumed one of the input classes is the masked latent pixel. The predicted classes of the unnoised image don’t contain a prediction for the masked pixel because the unnoised image cannot be masked.
Convert input (classes of latent pixels) to embeddings and apply positional embeddings.
Apply the Transformer blocks in the standard way.
Predict classes of unnoised image.
( num_attention_heads: int = 16attention_head_dim: int = 88in_channels: typing.Optional[int] = Noneout_channels: typing.Optional[int] = Nonenum_layers: int = 1dropout: float = 0.0norm_num_groups: int = 32cross_attention_dim: typing.Optional[int] = Noneattention_bias: bool = Falsesample_size: typing.Optional[int] = Nonenum_vector_embeds: typing.Optional[int] = Nonepatch_size: typing.Optional[int] = Noneactivation_fn: str = 'geglu'num_embeds_ada_norm: typing.Optional[int] = Noneuse_linear_projection: bool = Falseonly_cross_attention: bool = Falsedouble_self_attention: bool = Falseupcast_attention: bool = Falsenorm_type: str = 'layer_norm'norm_elementwise_affine: bool = Trueattention_type: str = 'default' )
Parameters
num_attention_heads (int, optional, defaults to 16) — The number of heads to use for multi-head attention.
attention_head_dim (int, optional, defaults to 88) — The number of channels in each head.
in_channels (int, optional) — The number of channels in the input and output (specify if the input is continuous).
num_layers (int, optional, defaults to 1) — The number of layers of Transformer blocks to use.
dropout (float, optional, defaults to 0.0) — The dropout probability to use.
cross_attention_dim (int, optional) — The number of encoder_hidden_states dimensions to use.
sample_size (int, optional) — The width of the latent images (specify if the input is discrete). This is fixed during training since it is used to learn a number of position embeddings.
num_vector_embeds (int, optional) — The number of classes of the vector embeddings of the latent pixels (specify if the input is discrete). Includes the class for the masked latent pixel.
activation_fn (str, optional, defaults to "geglu") — Activation function to use in feed-forward.
num_embeds_ada_norm ( int, optional) — The number of diffusion steps used during training. Pass if at least one of the norm_layers is AdaLayerNorm. This is fixed during training since it is used to learn a number of embeddings that are added to the hidden states.
During inference, you can denoise for up to but not more steps than num_embeds_ada_norm.
attention_bias (bool, optional) — Configure if the TransformerBlocks attention should contain a bias parameter.
A 2D Transformer model for image-like data.
forward
( hidden_states: Tensorencoder_hidden_states: typing.Optional[torch.Tensor] = Nonetimestep: typing.Optional[torch.LongTensor] = Noneclass_labels: typing.Optional[torch.LongTensor] = Nonecross_attention_kwargs: typing.Dict[str, typing.Any] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneencoder_attention_mask: typing.Optional[torch.Tensor] = Nonereturn_dict: bool = True )
Parameters
hidden_states (torch.LongTensor of shape (batch size, num latent pixels) if discrete, torch.FloatTensor of shape (batch size, channel, height, width) if continuous) — Input hidden_states.
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.
timestep ( torch.LongTensor, optional) — Used to indicate denoising step. Optional timestep to be applied as an embedding in AdaLayerNorm.
class_labels ( torch.LongTensor of shape (batch size, num classes), optional) — Used to indicate class labels conditioning. Optional class labels to be applied as an embedding in AdaLayerZeroNorm.
encoder_attention_mask ( torch.Tensor, optional) — Cross-attention mask applied to encoder_hidden_states. Two formats supported:
Mask (batch, sequence_length) True = keep, False = discard.
Bias (batch, 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.
( sample: FloatTensor )
Parameters
return_dict (bool, optional, defaults to True) — Whether or not to return a instead of a plain tuple.
The forward method.
sample (torch.FloatTensor of shape (batch_size, num_channels, height, width) or (batch size, num_vector_embeds - 1, num_latent_pixels) if is discrete) — The hidden states output conditioned on the encoder_hidden_states input. If discrete, returns probability distributions for the unnoised latent pixels.
The output of .