Stable Diffusion¶
Stable Diffusion은 텍스트 프롬프트로부터 이미지를 생성할 수 있는 텍스트-이미지 잠재 확산 모델입니다. RBLN NPU는 Optimum RBLN을 사용하여 Stable Diffusion 파이프라인을 가속화할 수 있습니다.
지원하는 파이프라인¶
Optimum RBLN은 여러 Stable Diffusion 파이프라인을 지원합니다:
- 텍스트-이미지 변환(Text-to-Image): 텍스트 프롬프트에서 이미지 생성
- 이미지-이미지 변환(Image-to-Image): 텍스트 프롬프트를 기반으로 기존 이미지 수정
- 인페인팅(Inpainting): 텍스트 프롬프트에 따라 이미지의 마스킹된 영역 채우기
주요 클래스¶
RBLNStableDiffusionPipeline
: Stable Diffusion의 텍스트-이미지 파이프라인RBLNStableDiffusionPipelineConfig
: 텍스트-이미지 파이프라인 설정RBLNStableDiffusionImg2ImgPipeline
: Stable Diffusion의 이미지-이미지 파이프라인RBLNStableDiffusionImg2ImgPipelineConfig
: 이미지-이미지 파이프라인 설정RBLNStableDiffusionInpaintPipeline
: Stable Diffusion의 인페인팅 파이프라인RBLNStableDiffusionInpaintPipelineConfig
: 인페인팅 파이프라인 설정
중요: Guidance Scale에 따른 배치 크기 설정¶
배치 크기와 Guidance Scale
Stable Diffusion을 guidance scale > 1.0으로 사용할 때(기본값은 7.5), classifier-free guidance 기법으로 인해 UNet의 실제 배치 크기가 실행 시 2배가 됩니다.
RBLN NPU는 정적 그래프 컴파일을 사용하므로, 컴파일 시 UNet의 배치 크기가 실행 시 배치 크기와 일치해야 합니다. 그렇지 않으면 추론 중에 오류가 발생합니다.
기본 동작¶
UNet의 배치 크기를 명시적으로 지정하지 않는 경우, Optimum RBLN은 다음과 같이 동작합니다:
- 기본 guidance scale(7.5)을 사용한다고 가정합니다
- 자동으로 UNet의 배치 크기를 파이프라인 배치 크기의 2배로 설정합니다
기본 guidance scale(1.0보다 큰 값)을 사용할 계획이라면, 이 자동 구성이 올바르게 작동합니다. 그러나 다른 guidance scale을 사용하거나 더 많은 제어가 필요한 경우에는 UNet의 배치 크기를 명시적으로 구성해야 합니다.
예시: UNet 배치 크기 명시적 설정¶
예시: Guidance Scale 1.0 사용¶
사용 예제¶
API 참조¶
Classes¶
RBLNStableDiffusionPipeline
¶
Bases: RBLNDiffusionMixin
, StableDiffusionPipeline
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 frommodel_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:
|
required |
export
|
bool
|
If True, takes a PyTorch model from |
False
|
model_save_dir
|
Optional[PathLike]
|
Directory to save the compiled model artifacts. Only used when |
None
|
rbln_config
|
Dict[str, Any]
|
Configuration options for RBLN compilation. Can include settings for specific submodules
such as |
{}
|
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 |
None
|
lora_weights_names
|
Optional[Union[str, List[str]]]
|
Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when |
None
|
lora_scales
|
Optional[Union[float, List[float]]]
|
Scaling factor(s) to apply to the LoRA adapter(s). Only used when |
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. |
RBLNStableDiffusionInpaintPipeline
¶
Bases: RBLNDiffusionMixin
, StableDiffusionInpaintPipeline
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 frommodel_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:
|
required |
export
|
bool
|
If True, takes a PyTorch model from |
False
|
model_save_dir
|
Optional[PathLike]
|
Directory to save the compiled model artifacts. Only used when |
None
|
rbln_config
|
Dict[str, Any]
|
Configuration options for RBLN compilation. Can include settings for specific submodules
such as |
{}
|
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 |
None
|
lora_weights_names
|
Optional[Union[str, List[str]]]
|
Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when |
None
|
lora_scales
|
Optional[Union[float, List[float]]]
|
Scaling factor(s) to apply to the LoRA adapter(s). Only used when |
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. |
RBLNStableDiffusionImg2ImgPipeline
¶
Bases: RBLNDiffusionMixin
, StableDiffusionImg2ImgPipeline
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 frommodel_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:
|
required |
export
|
bool
|
If True, takes a PyTorch model from |
False
|
model_save_dir
|
Optional[PathLike]
|
Directory to save the compiled model artifacts. Only used when |
None
|
rbln_config
|
Dict[str, Any]
|
Configuration options for RBLN compilation. Can include settings for specific submodules
such as |
{}
|
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 |
None
|
lora_weights_names
|
Optional[Union[str, List[str]]]
|
Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when |
None
|
lora_scales
|
Optional[Union[float, List[float]]]
|
Scaling factor(s) to apply to the LoRA adapter(s). Only used when |
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¶
RBLNStableDiffusionPipelineBaseConfig
¶
Bases: RBLNModelConfig
Functions¶
__init__(text_encoder=None, unet=None, vae=None, *, batch_size=None, img_height=None, img_width=None, sample_size=None, image_size=None, guidance_scale=None, **kwargs)
¶
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text_encoder
|
Optional[RBLNCLIPTextModelConfig]
|
Configuration for the text encoder component. Initialized as RBLNCLIPTextModelConfig if not provided. |
None
|
unet
|
Optional[RBLNUNet2DConditionModelConfig]
|
Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided. |
None
|
vae
|
Optional[RBLNAutoencoderKLConfig]
|
Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided. |
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
|
sample_size
|
Optional[Tuple[int, int]]
|
Spatial dimensions for the UNet model. |
None
|
image_size
|
Optional[Tuple[int, int]]
|
Alternative way to specify image dimensions. Cannot be used together with img_height/img_width. |
None
|
guidance_scale
|
Optional[float]
|
Scale for classifier-free guidance. Deprecated parameter. |
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.
RBLNStableDiffusionPipelineConfig
¶
Bases: RBLNStableDiffusionPipelineBaseConfig
Functions¶
__init__(text_encoder=None, unet=None, vae=None, *, batch_size=None, img_height=None, img_width=None, sample_size=None, image_size=None, guidance_scale=None, **kwargs)
¶
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text_encoder
|
Optional[RBLNCLIPTextModelConfig]
|
Configuration for the text encoder component. Initialized as RBLNCLIPTextModelConfig if not provided. |
None
|
unet
|
Optional[RBLNUNet2DConditionModelConfig]
|
Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided. |
None
|
vae
|
Optional[RBLNAutoencoderKLConfig]
|
Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided. |
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
|
sample_size
|
Optional[Tuple[int, int]]
|
Spatial dimensions for the UNet model. |
None
|
image_size
|
Optional[Tuple[int, int]]
|
Alternative way to specify image dimensions. Cannot be used together with img_height/img_width. |
None
|
guidance_scale
|
Optional[float]
|
Scale for classifier-free guidance. Deprecated parameter. |
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.
RBLNStableDiffusionImg2ImgPipelineConfig
¶
Bases: RBLNStableDiffusionPipelineBaseConfig
Functions¶
__init__(text_encoder=None, unet=None, vae=None, *, batch_size=None, img_height=None, img_width=None, sample_size=None, image_size=None, guidance_scale=None, **kwargs)
¶
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text_encoder
|
Optional[RBLNCLIPTextModelConfig]
|
Configuration for the text encoder component. Initialized as RBLNCLIPTextModelConfig if not provided. |
None
|
unet
|
Optional[RBLNUNet2DConditionModelConfig]
|
Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided. |
None
|
vae
|
Optional[RBLNAutoencoderKLConfig]
|
Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided. |
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
|
sample_size
|
Optional[Tuple[int, int]]
|
Spatial dimensions for the UNet model. |
None
|
image_size
|
Optional[Tuple[int, int]]
|
Alternative way to specify image dimensions. Cannot be used together with img_height/img_width. |
None
|
guidance_scale
|
Optional[float]
|
Scale for classifier-free guidance. Deprecated parameter. |
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.
RBLNStableDiffusionInpaintPipelineConfig
¶
Bases: RBLNStableDiffusionPipelineBaseConfig
Functions¶
__init__(text_encoder=None, unet=None, vae=None, *, batch_size=None, img_height=None, img_width=None, sample_size=None, image_size=None, guidance_scale=None, **kwargs)
¶
Parameters:
Name | Type | Description | Default |
---|---|---|---|
text_encoder
|
Optional[RBLNCLIPTextModelConfig]
|
Configuration for the text encoder component. Initialized as RBLNCLIPTextModelConfig if not provided. |
None
|
unet
|
Optional[RBLNUNet2DConditionModelConfig]
|
Configuration for the UNet model component. Initialized as RBLNUNet2DConditionModelConfig if not provided. |
None
|
vae
|
Optional[RBLNAutoencoderKLConfig]
|
Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided. |
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
|
sample_size
|
Optional[Tuple[int, int]]
|
Spatial dimensions for the UNet model. |
None
|
image_size
|
Optional[Tuple[int, int]]
|
Alternative way to specify image dimensions. Cannot be used together with img_height/img_width. |
None
|
guidance_scale
|
Optional[float]
|
Scale for classifier-free guidance. Deprecated parameter. |
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.