Kandinsky V2.2

Kandinsky V2.2 is a text-to-image latent diffusion model. RBLN NPUs can accelerate Kandinsky V2.2 pipelines using Optimum RBLN.

Supported Pipelines

Optimum RBLN supports several Kandinsky V2.2 pipelines:

• Text-to-Image: Generate images from text prompts (using Prior + Decoder)
• Image-to-Image: Modify existing images based on text prompts (using Prior + Img2Img Decoder)
• Inpainting: Fill masked regions of an image guided by text prompts (using Prior + Inpaint Decoder)

Key Classes

• RBLNKandinskyV22PriorPipeline: Pipeline for the Prior stage (Text/Image -> Image Embedding)
• RBLNKandinskyV22PriorPipelineConfig: Configuration for the Prior pipeline
• RBLNKandinskyV22Pipeline: Text-to-image Decoder pipeline (Image Embedding -> Image)
• RBLNKandinskyV22PipelineConfig: Configuration for the Text-to-image Decoder pipeline
• RBLNKandinskyV22Img2ImgPipeline: Image-to-image Decoder pipeline
• RBLNKandinskyV22Img2ImgPipelineConfig: Configuration for the Image-to-image Decoder pipeline
• RBLNKandinskyV22InpaintPipeline: Inpainting Decoder pipeline
• RBLNKandinskyV22InpaintPipelineConfig: Configuration for the Inpainting Decoder pipeline
• RBLNKandinskyV22CombinedPipeline: Combined pipeline (Prior + Text-to-Image Decoder)
• RBLNKandinskyV22CombinedPipelineConfig: Configuration for the Combined pipeline
• RBLNKandinskyV22Img2ImgCombinedPipeline: Combined pipeline (Prior + Image-to-Image Decoder)
• RBLNKandinskyV22Img2ImgCombinedPipelineConfig: Configuration for the Combined pipeline (Prior + Image-to-Image Decoder)
• RBLNKandinskyV22InpaintCombinedPipeline: Combined pipeline (Prior + Inpainting Decoder)
• RBLNKandinskyV22InpaintCombinedPipelineConfig: Configuration for the Combined pipeline (Prior + Inpainting Decoder)

Important: Batch Size Configuration for Guidance Scale

Batch Size and Guidance Scale

When using Kandinsky V2.2 with a guidance scale > 1.0 (the default), both the UNet's and Prior's effective batch sizes are doubled during runtime because of the classifier-free guidance technique.

Since RBLN NPU uses static graph compilation, these components' batch sizes at compilation time must match their runtime batch sizes, or you'll encounter errors during inference.

Default Behavior

By default, if you don't explicitly specify the UNet's or Prior's batch size, Optimum RBLN will:

• Assume you'll use the default guidance scale (which is > 1.0)
• Automatically set the UNet's and Prior's batch sizes to 2× your pipeline's batch size

If you plan to use the default guidance scale, this automatic configuration will work correctly. However, if you plan to use a different guidance scale or want more control, you should explicitly configure the batch sizes.

Example: Explicitly Setting Batch Sizes (Guidance Scale = 1.0)

If you plan to use a guidance scale of exactly 1.0 (which doesn't use classifier-free guidance), you should explicitly set the batch sizes to match your inference batch size:

Usage Examples

Option 1: Using Separate Prior and Decoder Pipelines

This approach gives you more control over the intermediate image embeddings:

import torch
from optimum.rbln import RBLNKandinskyV22PriorPipeline, RBLNKandinskyV22Pipeline

# Load and compile the Prior pipeline
prior_pipe = RBLNKandinskyV22PriorPipeline.from_pretrained(
    "kandinsky-community/kandinsky-2-2-prior",
    export=True  # Set to True for first-time compilation
)

# Load and compile the Decoder pipeline with specific image dimensions
pipe = RBLNKandinskyV22Pipeline.from_pretrained(
    "kandinsky-community/kandinsky-2-2-decoder",
    export=True,
    rbln_img_height=768,
    rbln_img_width=768
)

# Generate image embeddings from text prompt using the Prior
prompt = "a cute cat sitting on a beach, 4k photo"
generator = torch.manual_seed(42)  # For reproducible results
prior_output = prior_pipe(prompt, generator=generator)
image_emb = prior_output.image_embeds
negative_image_emb = prior_output.negative_image_embeds

# Generate the final image using the Decoder and the image embeddings
output = pipe(
    image_embeds=image_emb,
    negative_image_embeds=negative_image_emb,
    height=768,
    width=768,
    num_inference_steps=50,
    generator=generator
)

# Save the generated image
image = output.images[0]
image.save("generated_cat.png")

Option 2: Using Combined Pipeline

The combined pipeline integrates both Prior and Decoder into a single seamless workflow:

import torch
from optimum.rbln import RBLNKandinskyV22CombinedPipeline

# Load and compile the Combined pipeline
pipe = RBLNKandinskyV22CombinedPipeline.from_pretrained(
    "kandinsky-community/kandinsky-2-2-decoder",  # Main decoder model
    export=True,  # Set to True for first-time compilation
    rbln_img_height=768,
    rbln_img_width=768
)

# Generate and save an image
prompt = "a cute cat sitting on a beach, 4k photo"
generator = torch.manual_seed(42)  # For reproducible results
image = pipe(
    prompt, 
    height=768, 
    width=768, 
    num_inference_steps=50, 
    generator=generator
).images[0]

image.save("generated_cat.png")

API Reference

Classes

RBLNKandinskyV22PriorPipeline

Bases: RBLNDiffusionMixin, KandinskyV22PriorPipeline

RBLN wrapper for Kandinsky V2.2 Prior pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

RBLNKandinskyV22Pipeline

Bases: RBLNDiffusionMixin, KandinskyV22Pipeline

RBLN wrapper for Kandinsky V2.2 text-to-image pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

RBLNKandinskyV22Img2ImgPipeline

Bases: RBLNDiffusionMixin, KandinskyV22Img2ImgPipeline

RBLN wrapper for Kandinsky V2.2 image-to-image pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

RBLNKandinskyV22InpaintPipeline

Bases: RBLNDiffusionMixin, KandinskyV22InpaintPipeline

RBLN wrapper for Kandinsky V2.2 inpainting pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

RBLNKandinskyV22CombinedPipeline

Bases: RBLNDiffusionMixin, KandinskyV22CombinedPipeline

RBLN wrapper for Kandinsky V2.2 Combined (Prior + Text-to-Image Decoder) pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

RBLNKandinskyV22Img2ImgCombinedPipeline

Bases: RBLNDiffusionMixin, KandinskyV22Img2ImgCombinedPipeline

RBLN wrapper for Kandinsky V2.2 Combined (Prior + Image-to-Image Decoder) pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

RBLNKandinskyV22InpaintCombinedPipeline

Bases: RBLNDiffusionMixin, KandinskyV22InpaintCombinedPipeline

RBLN wrapper for Kandinsky V2.2 Combined (Prior + Inpainting Decoder) pipeline.

Functions

from_pretrained(model_id, *, export=False, model_save_dir=None, rbln_config={}, lora_ids=None, lora_weights_names=None, lora_scales=None, **kwargs) classmethod

Load a pretrained diffusion pipeline from a model checkpoint, with optional compilation for RBLN NPUs.

This method has two distinct operating modes:

• When export=True: Takes a PyTorch-based diffusion model, compiles it for RBLN NPUs, and loads the compiled model
• When export=False: Loads an already compiled RBLN model from model_id without recompilation

It supports various diffusion pipelines including Stable Diffusion, Kandinsky, ControlNet, and other diffusers-based models.

Parameters:

Name	Type	Description	Default
model_id	str	The model ID or path to the pretrained model to load. Can be either: A model ID from the Hugging Face Hub A local path to a saved model directory	required
export	bool	If True, takes a PyTorch model from model_id and compiles it for RBLN NPU execution. If False, loads an already compiled RBLN model from model_id without recompilation.	False
model_save_dir	Optional[PathLike]	Directory to save the compiled model artifacts. Only used when export=True. If not provided and export=True, a temporary directory is used.	None
rbln_config	Dict[str, Any]	Configuration options for RBLN compilation. Can include settings for specific submodules such as text_encoder, unet, and vae. Configuration can be tailored to the specific pipeline being compiled.	{}
lora_ids	Optional[Union[str, List[str]]]	LoRA adapter ID(s) to load and apply before compilation. LoRA weights are fused into the model weights during compilation. Only used when export=True.	None
lora_weights_names	Optional[Union[str, List[str]]]	Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.	None
lora_scales	Optional[Union[float, List[float]]]	Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.	None
**kwargs	Dict[str, Any]	Additional arguments to pass to the underlying diffusion pipeline constructor or the RBLN compilation process. These may include parameters specific to individual submodules or the particular diffusion pipeline being used.	{}

Returns:

Type	Description
Self	A compiled diffusion pipeline that can be used for inference on RBLN NPU. The returned object is an instance of the class that called this method, inheriting from RBLNDiffusionMixin.

Classes

RBLNKandinskyV22PipelineBaseConfig

Bases: RBLNModelConfig

Base configuration class for Kandinsky V2.2 decoder pipelines.

Functions

__init__(unet=None, movq=None, *, sample_size=None, batch_size=None, guidance_scale=None, image_size=None, img_height=None, img_width=None, **kwargs)

Parameters:

Name	Type	Description	Default
unet	Optional[RBLNUNet2DConditionModelConfig]	Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided.	None
movq	Optional[RBLNVQModelConfig]	Configuration for the MoVQ (VQ-GAN) model component. Initialized as RBLNVQModelConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

Raises:

Type	Description
ValueError	If both image_size and img_height/img_width are provided.

Note

When guidance_scale > 1.0, the UNet batch size is automatically doubled to accommodate classifier-free guidance.

RBLNKandinskyV22PipelineConfig

Bases: RBLNKandinskyV22PipelineBaseConfig

Configuration class for the Kandinsky V2.2 text-to-image decoder pipeline.

Functions

__init__(unet=None, movq=None, *, sample_size=None, batch_size=None, guidance_scale=None, image_size=None, img_height=None, img_width=None, **kwargs)

Parameters:

Name	Type	Description	Default
unet	Optional[RBLNUNet2DConditionModelConfig]	Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided.	None
movq	Optional[RBLNVQModelConfig]	Configuration for the MoVQ (VQ-GAN) model component. Initialized as RBLNVQModelConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

Raises:

Type	Description
ValueError	If both image_size and img_height/img_width are provided.

Note

When guidance_scale > 1.0, the UNet batch size is automatically doubled to accommodate classifier-free guidance.

RBLNKandinskyV22Img2ImgPipelineConfig

Bases: RBLNKandinskyV22PipelineBaseConfig

Configuration class for the Kandinsky V2.2 image-to-image decoder pipeline.

Functions

__init__(unet=None, movq=None, *, sample_size=None, batch_size=None, guidance_scale=None, image_size=None, img_height=None, img_width=None, **kwargs)

Parameters:

Name	Type	Description	Default
unet	Optional[RBLNUNet2DConditionModelConfig]	Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided.	None
movq	Optional[RBLNVQModelConfig]	Configuration for the MoVQ (VQ-GAN) model component. Initialized as RBLNVQModelConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

Raises:

Type	Description
ValueError	If both image_size and img_height/img_width are provided.

Note

When guidance_scale > 1.0, the UNet batch size is automatically doubled to accommodate classifier-free guidance.

RBLNKandinskyV22InpaintPipelineConfig

Bases: RBLNKandinskyV22PipelineBaseConfig

Configuration class for the Kandinsky V2.2 inpainting decoder pipeline.

Functions

__init__(unet=None, movq=None, *, sample_size=None, batch_size=None, guidance_scale=None, image_size=None, img_height=None, img_width=None, **kwargs)

Parameters:

Name	Type	Description	Default
unet	Optional[RBLNUNet2DConditionModelConfig]	Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided.	None
movq	Optional[RBLNVQModelConfig]	Configuration for the MoVQ (VQ-GAN) model component. Initialized as RBLNVQModelConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

Raises:

Type	Description
ValueError	If both image_size and img_height/img_width are provided.

Note

When guidance_scale > 1.0, the UNet batch size is automatically doubled to accommodate classifier-free guidance.

RBLNKandinskyV22PriorPipelineConfig

Bases: RBLNModelConfig

Configuration class for the Kandinsky V2.2 Prior pipeline.

Functions

__init__(text_encoder=None, image_encoder=None, prior=None, *, batch_size=None, guidance_scale=None, **kwargs)

Parameters:

Name	Type	Description	Default
text_encoder	Optional[RBLNCLIPTextModelWithProjectionConfig]	Configuration for the text encoder component. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.	None
image_encoder	Optional[RBLNCLIPVisionModelWithProjectionConfig]	Configuration for the image encoder component. Initialized as RBLNCLIPVisionModelWithProjectionConfig if not provided.	None
prior	Optional[RBLNPriorTransformerConfig]	Configuration for the prior transformer component. Initialized as RBLNPriorTransformerConfig if not provided.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

Note

When guidance_scale > 1.0, the prior batch size is automatically doubled to accommodate classifier-free guidance.

RBLNKandinskyV22CombinedPipelineBaseConfig

Bases: RBLNModelConfig

Base configuration class for Kandinsky V2.2 combined pipelines.

Functions

__init__(prior_pipe=None, decoder_pipe=None, *, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, guidance_scale=None, prior_prior=None, prior_image_encoder=None, prior_text_encoder=None, unet=None, movq=None, **kwargs)

Parameters:

Name	Type	Description	Default
prior_pipe	Optional[RBLNKandinskyV22PriorPipelineConfig]	Configuration for the prior pipeline. Initialized as RBLNKandinskyV22PriorPipelineConfig if not provided.	None
decoder_pipe	Optional[RBLNKandinskyV22PipelineConfig]	Configuration for the decoder pipeline. Initialized as RBLNKandinskyV22PipelineConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
prior_prior	Optional[RBLNPriorTransformerConfig]	Direct configuration for the prior transformer. Used if prior_pipe is not provided.	None
prior_image_encoder	Optional[RBLNCLIPVisionModelWithProjectionConfig]	Direct configuration for the image encoder. Used if prior_pipe is not provided.	None
prior_text_encoder	Optional[RBLNCLIPTextModelWithProjectionConfig]	Direct configuration for the text encoder. Used if prior_pipe is not provided.	None
unet	Optional[RBLNUNet2DConditionModelConfig]	Direct configuration for the UNet. Used if decoder_pipe is not provided.	None
movq	Optional[RBLNVQModelConfig]	Direct configuration for the MoVQ (VQ-GAN) model. Used if decoder_pipe is not provided.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

RBLNKandinskyV22CombinedPipelineConfig

Bases: RBLNKandinskyV22CombinedPipelineBaseConfig

Configuration class for the Kandinsky V2.2 combined text-to-image pipeline.

Functions

__init__(prior_pipe=None, decoder_pipe=None, *, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, guidance_scale=None, prior_prior=None, prior_image_encoder=None, prior_text_encoder=None, unet=None, movq=None, **kwargs)

Parameters:

Name	Type	Description	Default
prior_pipe	Optional[RBLNKandinskyV22PriorPipelineConfig]	Configuration for the prior pipeline. Initialized as RBLNKandinskyV22PriorPipelineConfig if not provided.	None
decoder_pipe	Optional[RBLNKandinskyV22PipelineConfig]	Configuration for the decoder pipeline. Initialized as RBLNKandinskyV22PipelineConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
prior_prior	Optional[RBLNPriorTransformerConfig]	Direct configuration for the prior transformer. Used if prior_pipe is not provided.	None
prior_image_encoder	Optional[RBLNCLIPVisionModelWithProjectionConfig]	Direct configuration for the image encoder. Used if prior_pipe is not provided.	None
prior_text_encoder	Optional[RBLNCLIPTextModelWithProjectionConfig]	Direct configuration for the text encoder. Used if prior_pipe is not provided.	None
unet	Optional[RBLNUNet2DConditionModelConfig]	Direct configuration for the UNet. Used if decoder_pipe is not provided.	None
movq	Optional[RBLNVQModelConfig]	Direct configuration for the MoVQ (VQ-GAN) model. Used if decoder_pipe is not provided.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

RBLNKandinskyV22InpaintCombinedPipelineConfig

Bases: RBLNKandinskyV22CombinedPipelineBaseConfig

Configuration class for the Kandinsky V2.2 combined inpainting pipeline.

Functions

__init__(prior_pipe=None, decoder_pipe=None, *, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, guidance_scale=None, prior_prior=None, prior_image_encoder=None, prior_text_encoder=None, unet=None, movq=None, **kwargs)

Parameters:

Name	Type	Description	Default
prior_pipe	Optional[RBLNKandinskyV22PriorPipelineConfig]	Configuration for the prior pipeline. Initialized as RBLNKandinskyV22PriorPipelineConfig if not provided.	None
decoder_pipe	Optional[RBLNKandinskyV22PipelineConfig]	Configuration for the decoder pipeline. Initialized as RBLNKandinskyV22PipelineConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
prior_prior	Optional[RBLNPriorTransformerConfig]	Direct configuration for the prior transformer. Used if prior_pipe is not provided.	None
prior_image_encoder	Optional[RBLNCLIPVisionModelWithProjectionConfig]	Direct configuration for the image encoder. Used if prior_pipe is not provided.	None
prior_text_encoder	Optional[RBLNCLIPTextModelWithProjectionConfig]	Direct configuration for the text encoder. Used if prior_pipe is not provided.	None
unet	Optional[RBLNUNet2DConditionModelConfig]	Direct configuration for the UNet. Used if decoder_pipe is not provided.	None
movq	Optional[RBLNVQModelConfig]	Direct configuration for the MoVQ (VQ-GAN) model. Used if decoder_pipe is not provided.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}

RBLNKandinskyV22Img2ImgCombinedPipelineConfig

Bases: RBLNKandinskyV22CombinedPipelineBaseConfig

Configuration class for the Kandinsky V2.2 combined image-to-image pipeline.

Functions

__init__(prior_pipe=None, decoder_pipe=None, *, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, guidance_scale=None, prior_prior=None, prior_image_encoder=None, prior_text_encoder=None, unet=None, movq=None, **kwargs)

Parameters:

Name	Type	Description	Default
prior_pipe	Optional[RBLNKandinskyV22PriorPipelineConfig]	Configuration for the prior pipeline. Initialized as RBLNKandinskyV22PriorPipelineConfig if not provided.	None
decoder_pipe	Optional[RBLNKandinskyV22PipelineConfig]	Configuration for the decoder pipeline. Initialized as RBLNKandinskyV22PipelineConfig if not provided.	None
sample_size	Optional[Tuple[int, int]]	Spatial dimensions for the UNet model.	None
image_size	Optional[Tuple[int, int]]	Dimensions for the generated images. Cannot be used together with img_height/img_width.	None
batch_size	Optional[int]	Batch size for inference, applied to all submodules.	None
img_height	Optional[int]	Height of the generated images.	None
img_width	Optional[int]	Width of the generated images.	None
guidance_scale	Optional[float]	Scale for classifier-free guidance.	None
prior_prior	Optional[RBLNPriorTransformerConfig]	Direct configuration for the prior transformer. Used if prior_pipe is not provided.	None
prior_image_encoder	Optional[RBLNCLIPVisionModelWithProjectionConfig]	Direct configuration for the image encoder. Used if prior_pipe is not provided.	None
prior_text_encoder	Optional[RBLNCLIPTextModelWithProjectionConfig]	Direct configuration for the text encoder. Used if prior_pipe is not provided.	None
unet	Optional[RBLNUNet2DConditionModelConfig]	Direct configuration for the UNet. Used if decoder_pipe is not provided.	None
movq	Optional[RBLNVQModelConfig]	Direct configuration for the MoVQ (VQ-GAN) model. Used if decoder_pipe is not provided.	None
**kwargs	Dict[str, Any]	Additional arguments passed to the parent RBLNModelConfig.	{}