Python module

max.pipelines.architectures.flux1_modulev3

FLUX.1 diffusion architecture for image generation.

`FluxArchConfig`

class max.pipelines.architectures.flux1_modulev3.FluxArchConfig(*, max_seq_len=77, secondary_max_seq_len=512)

source

Pipeline-level config for Flux1 (implements ArchConfig; no KV cache).

Parameters:

get_max_seq_len()

source

Returns the maximum sequence length for the primary tokenizer.

classmethod initialize(pipeline_config, model_config=None)

source

Initialize the config from a PipelineConfig.

Parameters:

pipeline_config (PipelineConfig) – The pipeline configuration.
model_config (MAXModelConfig | None) – The model configuration to read from. When None (the default), pipeline_config.model is used. Pass an explicit config (e.g. pipeline_config.draft_model) to initialize the arch config for a different model.

Return type:

Self

max_seq_len: int = 77

source

secondary_max_seq_len: int = 512

source