콘텐츠로 이동

Stable Diffusion 3

Stable Diffusion 3(SD3)는 Stability AI의 텍스트-이미지 모델로, Multimodal Diffusion Transformer(MMDiT) 아키텍처를 기반으로 합니다. 여러 피사체, 공간 관계, 다양한 텍스트 스타일을 포함하는 복잡한 프롬프트를 처리하는 데 뛰어납니다. SD3는 향상된 프롬프트 해석을 위해 세 가지 고유한 텍스트 인코더(CLIP-L, OpenCLIP-G, T5-XXL)를 사용합니다. RBLN Optimum을 사용하면 RBLN NPU에서 Stable Diffusion 3 파이프라인을 가속할 수 있습니다.

지원 NPU

Diffusers 파이프라인은 ATOM™에서만 실행됩니다. Rebel100™은 아직 지원하지 않습니다.

지원하는 파이프라인

RBLN Optimum은 여러 Stable Diffusion 3 파이프라인을 지원합니다:

파이프라인 클래스 설명
텍스트-이미지 변환(Text-to-Image) RBLNStableDiffusion3Pipeline 텍스트 프롬프트에서 이미지 생성.
이미지-이미지 변환(Image-to-Image) RBLNStableDiffusion3Img2ImgPipeline 텍스트 프롬프트를 기반으로 기존 이미지 수정.
인페인팅(Inpainting) RBLNStableDiffusion3InpaintPipeline 텍스트 프롬프트에 따라 이미지의 마스킹된 영역 채우기.

중요: Guidance Scale에 따른 배치 크기 설정

배치 크기와 Guidance Scale

Stable Diffusion 3을 guidance scale > 1.0으로 사용할 때(기본값은 7.0), classifier-free guidance가 적용되어 MMDiT Transformer의 실제 배치 크기가 실행 시 2배가 됩니다.

RBLN NPU는 정적 그래프 컴파일을 사용하므로, 컴파일 시 Transformer의 배치 크기가 실행 시 배치 크기와 일치해야 합니다. 그렇지 않으면 컴파일된 그래프를 실행할 수 없어 추론이 실패합니다.

기본 동작

Transformer의 배치 크기를 명시적으로 지정하지 않는 경우, RBLN Optimum은 다음과 같이 동작합니다:

  • guidance scale > 1.0을 사용한다고 가정합니다.
  • 자동으로 Transformer의 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 Transformer의 배치 크기를 명시적으로 구성해야 합니다.

예시: Transformer 배치 크기 명시적 설정(guidance_scale > 1.0)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
import torch
from optimum.rbln import (
    RBLNStableDiffusion3PipelineConfig,
    RBLNStableDiffusion3Pipeline,
)

# guidance_scale > 1.0인 경우 (예: 기본값 7.0)
# Transformer의 배치 크기를 원하는 추론 배치 크기의 2배로 설정
config = RBLNStableDiffusion3PipelineConfig(
    # text_encoder_3 (~8.7 GiB) does not fit on the same ATOM™ as the rest; give it a second device
    text_encoder={"device": 0},
    text_encoder_2={"device": 0},
    text_encoder_3={"device": 1},
    vae={"device": 0},
    batch_size=1,  # 추론 배치 크기
    height=1024,
    width=1024,
    # Transformer의 배치 크기를 2배로 구성
    transformer=dict(batch_size=2, device=0),  # Transformer 배치 크기를 2배로 설정
)

pipe = RBLNStableDiffusion3Pipeline.from_pretrained(
    "stabilityai/stable-diffusion-3-medium-diffusers",
    torch_dtype=torch.float16,
    export=True,
    rbln_config=config,
)

# 기본 guidance_scale (예: 7.0)로 일반 추론
prompt = "A cat riding a bike through a field of sunflowers, realistic photo."
image = pipe(prompt).images[0]  # 배치 크기 1

# 생성된 이미지 저장
image.save("sd3_generated_image.png")
print("이미지가 sd3_generated_image.png로 저장되었습니다")

예시: Guidance 비활성화(guidance_scale = 0.0)

guidance_scale을 정확히 0.0으로 사용할 계획이라면(classifier-free guidance 비활성화), Transformer 배치 크기를 추론 배치 크기와 일치하도록 명시적으로 설정합니다:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
import torch
from optimum.rbln import (
    RBLNStableDiffusion3PipelineConfig,
    RBLNStableDiffusion3Pipeline,
)

config = RBLNStableDiffusion3PipelineConfig(
    # text_encoder_3 (~8.7 GiB) does not fit on the same ATOM™ as the rest; give it a second device
    text_encoder={"device": 0},
    text_encoder_2={"device": 0},
    text_encoder_3={"device": 1},
    vae={"device": 0},
    height=1024,
    width=1024,
    # Transformer 배치 크기가 추론 배치 크기와 일치
    transformer=dict(batch_size=1, device=0),
)

pipe = RBLNStableDiffusion3Pipeline.from_pretrained(
    "stabilityai/stable-diffusion-3-medium-diffusers",
    torch_dtype=torch.float16,
    export=True,
    rbln_config=config,
)

# 추론 시 반드시 guidance_scale=0.0 사용
prompt = "A cat riding a bike through a field of sunflowers, realistic photo."
image = pipe(prompt, guidance_scale=0.0).images[0]

# 생성된 이미지 저장
image.save("sd3_no_guidance_image.png")
print("이미지가 sd3_no_guidance_image.png로 저장되었습니다")

사용 예제(텍스트-이미지)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
import torch
from optimum.rbln import (
    RBLNStableDiffusion3Pipeline,
    RBLNStableDiffusion3PipelineConfig,
)

# 구성 객체 생성 (선택 사항, 기본값 사용 가능)
config = RBLNStableDiffusion3PipelineConfig(
    height=1024,
    width=1024,
    # text_encoder_3 (~8.7 GiB) does not fit on the same ATOM™ as the rest; give it a second device
    text_encoder={"device": 0},
    text_encoder_2={"device": 0},
    text_encoder_3={"device": 1},
    vae={"device": 0},
    transformer={"device": 0},
)

# RBLN NPU용 SD3 모델 로드 및 컴파일
pipe = RBLNStableDiffusion3Pipeline.from_pretrained(
    "stabilityai/stable-diffusion-3-medium-diffusers",
    torch_dtype=torch.float16,
    export=True,
    rbln_config=config,  # 생성한 경우 config 전달
)

# 이미지 생성
prompt = "A cat riding a bike through a field of sunflowers, realistic photo."
image = pipe(prompt).images[0]

# 생성된 이미지 저장
image.save("sd3_cat_bike.png")
print("이미지가 sd3_cat_bike.png로 저장되었습니다")

API 참조

RBLNStableDiffusion3Pipeline

Bases: RBLNDiffusionMixin, StableDiffusion3Pipeline

RBLN-accelerated implementation of Stable Diffusion 3 pipeline for advanced text-to-image generation.

This pipeline compiles Stable Diffusion 3 models to run efficiently on RBLN NPUs, enabling high-performance inference with improved text understanding, enhanced image quality, and superior prompt adherence.

from_pretrained(model_id, *, export=None, model_save_dir=None, rbln_config=None, 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 HuggingFace Hub
  • A local path to a saved model directory
 required
export bool | None

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.

 None
model_save_dir str | PathLike[str] | None

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] | RBLNModelConfig | None

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.

 None
lora_ids str | list[str] | None

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 str | list[str] | None

Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.

 None
lora_scales float | list[float] | None

Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.

 None
kwargs 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 or loaded 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.

RBLNStableDiffusion3Img2ImgPipeline

Bases: RBLNDiffusionMixin, StableDiffusion3Img2ImgPipeline

RBLN-accelerated implementation of Stable Diffusion 3 pipeline for advanced image-to-image generation.

This pipeline compiles Stable Diffusion 3 models to run efficiently on RBLN NPUs, enabling high-performance inference for transforming input images with superior text understanding and enhanced visual quality.

from_pretrained(model_id, *, export=None, model_save_dir=None, rbln_config=None, 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 HuggingFace Hub
  • A local path to a saved model directory
 required
export bool | None

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.

 None
model_save_dir str | PathLike[str] | None

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] | RBLNModelConfig | None

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.

 None
lora_ids str | list[str] | None

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 str | list[str] | None

Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.

 None
lora_scales float | list[float] | None

Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.

 None
kwargs 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 or loaded 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.

RBLNStableDiffusion3InpaintPipeline

Bases: RBLNDiffusionMixin, StableDiffusion3InpaintPipeline

RBLN-accelerated implementation of Stable Diffusion 3 pipeline for advanced image inpainting.

This pipeline compiles Stable Diffusion 3 models to run efficiently on RBLN NPUs, enabling high-performance inference for filling masked regions with superior text understanding and seamless content generation.

from_pretrained(model_id, *, export=None, model_save_dir=None, rbln_config=None, 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 HuggingFace Hub
  • A local path to a saved model directory
 required
export bool | None

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.

 None
model_save_dir str | PathLike[str] | None

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] | RBLNModelConfig | None

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.

 None
lora_ids str | list[str] | None

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 str | list[str] | None

Names of specific LoRA weight files to load, corresponding to lora_ids. Only used when export=True.

 None
lora_scales float | list[float] | None

Scaling factor(s) to apply to the LoRA adapter(s). Only used when export=True.

 None
kwargs 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 or loaded 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.

RBLNStableDiffusion3PipelineBaseConfig

Bases: RBLNModelConfig

__init__(transformer=None, text_encoder=None, text_encoder_2=None, text_encoder_3=None, vae=None, *, max_seq_len=None, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, height=None, width=None, guidance_scale=None, **kwargs)

Parameters:

Name Type Description Default
transformer RBLNSD3Transformer2DModelConfig | None

Configuration for the transformer model component. Initialized as RBLNSD3Transformer2DModelConfig if not provided.

 None
text_encoder RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the primary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_2 RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the secondary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_3 RBLNT5EncoderModelConfig | None

Configuration for the tertiary text encoder. Initialized as RBLNT5EncoderModelConfig if not provided.

 None
vae RBLNAutoencoderKLConfig | None

Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided.

 None
max_seq_len int | None

Maximum sequence length for text inputs. Defaults to 256.

 None
sample_size tuple[int, int] | None

Spatial dimensions for the transformer model.

 None
image_size tuple[int, int] | None

Dimensions for the generated images. Cannot be used together with img_height/img_width.

 None
batch_size int | None

Batch size for inference, applied to all submodules.

 None
img_height int | None

Height of the generated images.

 None
img_width int | None

Width of the generated images.

 None
height int | None

Height of the generated images.

 None
width int | None

Width of the generated images.

 None
guidance_scale float | None

Scale for classifier-free guidance.

 None
kwargs 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 transformer batch size is automatically doubled to accommodate classifier-free guidance.

from_pretrained(path, rbln_config=None, return_unused_kwargs=False, **kwargs) classmethod

Load a RBLNModelConfig from a path.

Parameters:

Name Type Description Default
path str

Path to the RBLNModelConfig file or directory containing the config file.

 required
rbln_config dict[str, Any] | None

Additional configuration to override.

 None
return_unused_kwargs bool

Whether to return unused kwargs.

 False
kwargs Any

Additional keyword arguments to override configuration values. Keys starting with 'rbln_' will have the prefix removed and be used to update the configuration.

 {}

Returns:

Name Type Description
RBLNModelConfig Self | tuple[Self, dict[str, Any]]

The loaded configuration instance.
Note

This method loads the configuration from the specified path and applies any provided overrides. If the loaded configuration class doesn't match the expected class, a warning will be logged.

Examples:

1
config = RBLNResNetForImageClassificationConfig.from_pretrained("/path/to/model")

initialize_submodule_config(submodule_config=None, force_kwargs=False, **kwargs)

Build the config of a submodule from the parent's runtime options and kwargs.

Returns the submodule config instance when its class is known (an instance was passed, or cls_name was given), otherwise the merged kwargs dict for the model class to resolve later.

get_load_overrides()

Extract the options that must cross the load boundary, for passing to from_pretrained.

Returns the explicitly-set runtime options (device, device_map, create_runtimes, activate_profiler, timeout) and the subclass-specific non-save attributes (load-behavior flags such as _load_visual_runtime, which are never serialized so they can only travel with the caller), recursively including submodules. Compile-time attributes are not included: when loading a compiled model they come from the artifact's rbln_config.json, and this dict is merged on top of it.

Returns:

Type Description
dict[str, Any]

Dictionary of load overrides, keyed like the rbln_config dict accepted by from_pretrained.

RBLNStableDiffusion3PipelineConfig

Bases: RBLNStableDiffusion3PipelineBaseConfig

Config for SD3 Text2Img Pipeline

initialize_submodule_config(submodule_config=None, force_kwargs=False, **kwargs)

Build the config of a submodule from the parent's runtime options and kwargs.

Returns the submodule config instance when its class is known (an instance was passed, or cls_name was given), otherwise the merged kwargs dict for the model class to resolve later.

__init__(transformer=None, text_encoder=None, text_encoder_2=None, text_encoder_3=None, vae=None, *, max_seq_len=None, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, height=None, width=None, guidance_scale=None, **kwargs)

Parameters:

Name Type Description Default
transformer RBLNSD3Transformer2DModelConfig | None

Configuration for the transformer model component. Initialized as RBLNSD3Transformer2DModelConfig if not provided.

 None
text_encoder RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the primary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_2 RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the secondary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_3 RBLNT5EncoderModelConfig | None

Configuration for the tertiary text encoder. Initialized as RBLNT5EncoderModelConfig if not provided.

 None
vae RBLNAutoencoderKLConfig | None

Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided.

 None
max_seq_len int | None

Maximum sequence length for text inputs. Defaults to 256.

 None
sample_size tuple[int, int] | None

Spatial dimensions for the transformer model.

 None
image_size tuple[int, int] | None

Dimensions for the generated images. Cannot be used together with img_height/img_width.

 None
batch_size int | None

Batch size for inference, applied to all submodules.

 None
img_height int | None

Height of the generated images.

 None
img_width int | None

Width of the generated images.

 None
height int | None

Height of the generated images.

 None
width int | None

Width of the generated images.

 None
guidance_scale float | None

Scale for classifier-free guidance.

 None
kwargs 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 transformer batch size is automatically doubled to accommodate classifier-free guidance.

get_load_overrides()

Extract the options that must cross the load boundary, for passing to from_pretrained.

Returns the explicitly-set runtime options (device, device_map, create_runtimes, activate_profiler, timeout) and the subclass-specific non-save attributes (load-behavior flags such as _load_visual_runtime, which are never serialized so they can only travel with the caller), recursively including submodules. Compile-time attributes are not included: when loading a compiled model they come from the artifact's rbln_config.json, and this dict is merged on top of it.

Returns:

Type Description
dict[str, Any]

Dictionary of load overrides, keyed like the rbln_config dict accepted by from_pretrained.

from_pretrained(path, rbln_config=None, return_unused_kwargs=False, **kwargs) classmethod

Load a RBLNModelConfig from a path.

Parameters:

Name Type Description Default
path str

Path to the RBLNModelConfig file or directory containing the config file.

 required
rbln_config dict[str, Any] | None

Additional configuration to override.

 None
return_unused_kwargs bool

Whether to return unused kwargs.

 False
kwargs Any

Additional keyword arguments to override configuration values. Keys starting with 'rbln_' will have the prefix removed and be used to update the configuration.

 {}

Returns:

Name Type Description
RBLNModelConfig Self | tuple[Self, dict[str, Any]]

The loaded configuration instance.
Note

This method loads the configuration from the specified path and applies any provided overrides. If the loaded configuration class doesn't match the expected class, a warning will be logged.

Examples:

1
config = RBLNResNetForImageClassificationConfig.from_pretrained("/path/to/model")

RBLNStableDiffusion3Img2ImgPipelineConfig

Bases: RBLNStableDiffusion3PipelineBaseConfig

Config for SD3 Img2Img Pipeline

initialize_submodule_config(submodule_config=None, force_kwargs=False, **kwargs)

Build the config of a submodule from the parent's runtime options and kwargs.

Returns the submodule config instance when its class is known (an instance was passed, or cls_name was given), otherwise the merged kwargs dict for the model class to resolve later.

__init__(transformer=None, text_encoder=None, text_encoder_2=None, text_encoder_3=None, vae=None, *, max_seq_len=None, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, height=None, width=None, guidance_scale=None, **kwargs)

Parameters:

Name Type Description Default
transformer RBLNSD3Transformer2DModelConfig | None

Configuration for the transformer model component. Initialized as RBLNSD3Transformer2DModelConfig if not provided.

 None
text_encoder RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the primary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_2 RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the secondary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_3 RBLNT5EncoderModelConfig | None

Configuration for the tertiary text encoder. Initialized as RBLNT5EncoderModelConfig if not provided.

 None
vae RBLNAutoencoderKLConfig | None

Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided.

 None
max_seq_len int | None

Maximum sequence length for text inputs. Defaults to 256.

 None
sample_size tuple[int, int] | None

Spatial dimensions for the transformer model.

 None
image_size tuple[int, int] | None

Dimensions for the generated images. Cannot be used together with img_height/img_width.

 None
batch_size int | None

Batch size for inference, applied to all submodules.

 None
img_height int | None

Height of the generated images.

 None
img_width int | None

Width of the generated images.

 None
height int | None

Height of the generated images.

 None
width int | None

Width of the generated images.

 None
guidance_scale float | None

Scale for classifier-free guidance.

 None
kwargs 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 transformer batch size is automatically doubled to accommodate classifier-free guidance.

get_load_overrides()

Extract the options that must cross the load boundary, for passing to from_pretrained.

Returns the explicitly-set runtime options (device, device_map, create_runtimes, activate_profiler, timeout) and the subclass-specific non-save attributes (load-behavior flags such as _load_visual_runtime, which are never serialized so they can only travel with the caller), recursively including submodules. Compile-time attributes are not included: when loading a compiled model they come from the artifact's rbln_config.json, and this dict is merged on top of it.

Returns:

Type Description
dict[str, Any]

Dictionary of load overrides, keyed like the rbln_config dict accepted by from_pretrained.

from_pretrained(path, rbln_config=None, return_unused_kwargs=False, **kwargs) classmethod

Load a RBLNModelConfig from a path.

Parameters:

Name Type Description Default
path str

Path to the RBLNModelConfig file or directory containing the config file.

 required
rbln_config dict[str, Any] | None

Additional configuration to override.

 None
return_unused_kwargs bool

Whether to return unused kwargs.

 False
kwargs Any

Additional keyword arguments to override configuration values. Keys starting with 'rbln_' will have the prefix removed and be used to update the configuration.

 {}

Returns:

Name Type Description
RBLNModelConfig Self | tuple[Self, dict[str, Any]]

The loaded configuration instance.
Note

This method loads the configuration from the specified path and applies any provided overrides. If the loaded configuration class doesn't match the expected class, a warning will be logged.

Examples:

1
config = RBLNResNetForImageClassificationConfig.from_pretrained("/path/to/model")

RBLNStableDiffusion3InpaintPipelineConfig

Bases: RBLNStableDiffusion3PipelineBaseConfig

Config for SD3 Inpainting Pipeline

initialize_submodule_config(submodule_config=None, force_kwargs=False, **kwargs)

Build the config of a submodule from the parent's runtime options and kwargs.

Returns the submodule config instance when its class is known (an instance was passed, or cls_name was given), otherwise the merged kwargs dict for the model class to resolve later.

__init__(transformer=None, text_encoder=None, text_encoder_2=None, text_encoder_3=None, vae=None, *, max_seq_len=None, sample_size=None, image_size=None, batch_size=None, img_height=None, img_width=None, height=None, width=None, guidance_scale=None, **kwargs)

Parameters:

Name Type Description Default
transformer RBLNSD3Transformer2DModelConfig | None

Configuration for the transformer model component. Initialized as RBLNSD3Transformer2DModelConfig if not provided.

 None
text_encoder RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the primary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_2 RBLNCLIPTextModelWithProjectionConfig | None

Configuration for the secondary text encoder. Initialized as RBLNCLIPTextModelWithProjectionConfig if not provided.

 None
text_encoder_3 RBLNT5EncoderModelConfig | None

Configuration for the tertiary text encoder. Initialized as RBLNT5EncoderModelConfig if not provided.

 None
vae RBLNAutoencoderKLConfig | None

Configuration for the VAE model component. Initialized as RBLNAutoencoderKLConfig if not provided.

 None
max_seq_len int | None

Maximum sequence length for text inputs. Defaults to 256.

 None
sample_size tuple[int, int] | None

Spatial dimensions for the transformer model.

 None
image_size tuple[int, int] | None

Dimensions for the generated images. Cannot be used together with img_height/img_width.

 None
batch_size int | None

Batch size for inference, applied to all submodules.

 None
img_height int | None

Height of the generated images.

 None
img_width int | None

Width of the generated images.

 None
height int | None

Height of the generated images.

 None
width int | None

Width of the generated images.

 None
guidance_scale float | None

Scale for classifier-free guidance.

 None
kwargs 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 transformer batch size is automatically doubled to accommodate classifier-free guidance.

get_load_overrides()

Extract the options that must cross the load boundary, for passing to from_pretrained.

Returns the explicitly-set runtime options (device, device_map, create_runtimes, activate_profiler, timeout) and the subclass-specific non-save attributes (load-behavior flags such as _load_visual_runtime, which are never serialized so they can only travel with the caller), recursively including submodules. Compile-time attributes are not included: when loading a compiled model they come from the artifact's rbln_config.json, and this dict is merged on top of it.

Returns:

Type Description
dict[str, Any]

Dictionary of load overrides, keyed like the rbln_config dict accepted by from_pretrained.

from_pretrained(path, rbln_config=None, return_unused_kwargs=False, **kwargs) classmethod

Load a RBLNModelConfig from a path.

Parameters:

Name Type Description Default
path str

Path to the RBLNModelConfig file or directory containing the config file.

 required
rbln_config dict[str, Any] | None

Additional configuration to override.

 None
return_unused_kwargs bool

Whether to return unused kwargs.

 False
kwargs Any

Additional keyword arguments to override configuration values. Keys starting with 'rbln_' will have the prefix removed and be used to update the configuration.

 {}

Returns:

Name Type Description
RBLNModelConfig Self | tuple[Self, dict[str, Any]]

The loaded configuration instance.
Note

This method loads the configuration from the specified path and applies any provided overrides. If the loaded configuration class doesn't match the expected class, a warning will be logged.

Examples:

1
config = RBLNResNetForImageClassificationConfig.from_pretrained("/path/to/model")