Diffusers

You are viewing main version, which requires installation from source. If you'd like regular pip install, checkout the latest stable version (v0.42.1).

Join the Hugging Face community

and get access to the augmented documentation experience

Collaborate on models, datasets and Spaces

Faster examples with accelerated inference

Switch between documentation themes

to get started

ErnieImageTransformer2DModel

Um modelo Transformer para dados semelhantes a imagens do ERNIE-Image.

Um modelo Transformer para dados semelhantes a imagens do ERNIE-Image-Turbo.

ErnieImageTransformer2DModel

classe diffusers.ErnieImageTransformer2DModel

< source >

( hidden_size: int = 3072num_attention_heads: int = 24num_layers: int = 24ffn_hidden_size: int = 8192in_channels: int = 128out_channels: int = 128patch_size: int = 1text_in_dim: int = 2560rope_theta: int = 256rope_axes_dim: typing.Tuple[int, int, int] = (32, 48, 48)eps: float = 1e-06qk_layernorm: bool = True )

forward

< source >

( hidden_states: Tensortimestep: Tensortext_bth: Tensortext_lens: Tensorreturn_dict: bool = True )

Parâmetros

hidden_states (torch.Tensor de forma (batch_size, in_channels, height, width)) — hidden_states de entrada.
timestep (torch.LongTensor) — Usado para indicar o passo de desruido.
text_bth (torch.Tensor) — Embeddings de texto condicionais (embeddings calculados a partir das condições de entrada, como prompts) a serem usados, com forma (batch_size, text_length, embed_dims).
text_lens (torch.Tensor) — Comprimentos das sequências de texto por amostra usados para construir a máscara de atenção.
return_dict (bool, opcional, padrão True) — Se deve ou não retornar um ~models.transformer_2d.Transformer2DModelOutput em vez de uma tupla simples.

O método forward do ErnieImageTransformer2DModel.

Update on GitHub

←EasyAnimateTransformer3DModel Flux2Transformer2DModel→