diffusers/main/en/api/models/flux2_transformer.md

# Flux2Transformer2DModel

A Transformer model for image-like data from [Flux2](https://hf.co/black-forest-labs/FLUX.2-dev).

## Flux2Transformer2DModel[[diffusers.Flux2Transformer2DModel]]

#### diffusers.Flux2Transformer2DModel[[diffusers.Flux2Transformer2DModel]]

[Source](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/transformers/transformer_flux2.py#L1040)

The Transformer model introduced in Flux 2.

Reference: https://blackforestlabs.ai/announcing-black-forest-labs/

forwarddiffusers.Flux2Transformer2DModel.forwardhttps://github.com/huggingface/diffusers/blob/main/src/diffusers/models/transformers/transformer_flux2.py#L1178[{"name": "hidden_states", "val": ": Tensor"}, {"name": "encoder_hidden_states", "val": ": Tensor = None"}, {"name": "timestep", "val": ": LongTensor = None"}, {"name": "img_ids", "val": ": Tensor = None"}, {"name": "txt_ids", "val": ": Tensor = None"}, {"name": "guidance", "val": ": Tensor = None"}, {"name": "joint_attention_kwargs", "val": ": dict[str, typing.Any] | None = None"}, {"name": "return_dict", "val": ": bool = True"}, {"name": "kv_cache", "val": ": Flux2KVCache | None = None"}, {"name": "kv_cache_mode", "val": ": str | None = None"}, {"name": "num_ref_tokens", "val": ": int = 0"}, {"name": "ref_fixed_timestep", "val": ": float = 0.0"}]- **hidden_states** (`torch.Tensor` of shape `(batch_size, image_sequence_length, in_channels)`) --
  Input `hidden_states`.
- **encoder_hidden_states** (`torch.Tensor` of shape `(batch_size, text_sequence_length, joint_attention_dim)`) --
  Conditional embeddings (embeddings computed from the input conditions such as prompts) to use.
- **timestep** (`torch.LongTensor`) --
  Used to indicate denoising step.
- **joint_attention_kwargs** (`dict`, *optional*) --
  A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
  `self.processor` in
  [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
- **return_dict** (`bool`, *optional*, defaults to `True`) --
  Whether or not to return a `~models.transformer_2d.Transformer2DModelOutput` instead of a plain
  tuple.
- **kv_cache** (`Flux2KVCache`, *optional*) --
  KV cache for reference image tokens. When `kv_cache_mode` is "extract", a new cache is created and
  returned. When "cached", the provided cache is used to inject ref K/V during attention.
- **kv_cache_mode** (`str`, *optional*) --
  One of "extract" (first step with ref tokens) or "cached" (subsequent steps using cached ref K/V). When
  `None`, standard forward pass without KV caching.
- **num_ref_tokens** (`int`, defaults to `0`) --
  Number of reference image tokens prepended to `hidden_states` (only used when
  `kv_cache_mode="extract"`).
- **ref_fixed_timestep** (`float`, defaults to `0.0`) --
  Fixed timestep for reference token modulation (only used when `kv_cache_mode="extract"`).0If `return_dict` is True, an `~models.transformer_2d.Transformer2DModelOutput` is returned, otherwise a
`tuple` where the first element is the sample tensor. When `kv_cache_mode="extract"`, also returns the
populated `Flux2KVCache`.

The [Flux2Transformer2DModel](/docs/diffusers/main/en/api/models/flux2_transformer#diffusers.Flux2Transformer2DModel) forward method.

**Parameters:**

patch_size (`int`, defaults to `1`) : Patch size to turn the input data into small patches.

in_channels (`int`, defaults to `128`) : The number of channels in the input.

out_channels (`int`, *optional*, defaults to `None`) : The number of channels in the output. If not specified, it defaults to `in_channels`.

num_layers (`int`, defaults to `8`) : The number of layers of dual stream DiT blocks to use.

num_single_layers (`int`, defaults to `48`) : The number of layers of single stream DiT blocks to use.

attention_head_dim (`int`, defaults to `128`) : The number of dimensions to use for each attention head.

num_attention_heads (`int`, defaults to `48`) : The number of attention heads to use.

joint_attention_dim (`int`, defaults to `15360`) : The number of dimensions to use for the joint attention (embedding/channel dimension of `encoder_hidden_states`).

pooled_projection_dim (`int`, defaults to `768`) : The number of dimensions to use for the pooled projection.

guidance_embeds (`bool`, defaults to `True`) : Whether to use guidance embeddings for guidance-distilled variant of the model.

axes_dims_rope (`tuple[int]`, defaults to `(32, 32, 32, 32)`) : The dimensions to use for the rotary positional embeddings.

**Returns:**

If `return_dict` is True, an `~models.transformer_2d.Transformer2DModelOutput` is returned, otherwise a
`tuple` where the first element is the sample tensor. When `kv_cache_mode="extract"`, also returns the
populated `Flux2KVCache`.

## Flux2Transformer2DModelOutput[[diffusers.models.transformers.transformer_flux2.Flux2Transformer2DModelOutput]]

#### diffusers.models.transformers.transformer_flux2.Flux2Transformer2DModelOutput[[diffusers.models.transformers.transformer_flux2.Flux2Transformer2DModelOutput]]

[Source](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/transformers/transformer_flux2.py#L44)

The output of [Flux2Transformer2DModel](/docs/diffusers/main/en/api/models/flux2_transformer#diffusers.Flux2Transformer2DModel).

**Parameters:**

sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)`) : The hidden states output conditioned on the `encoder_hidden_states` input.

kv_cache (`Flux2KVCache`, *optional*) : The populated KV cache for reference image tokens. Only returned when `kv_cache_mode="extract"`.