Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'main' into fix/inpaint_gen

Browse Source
This commit is contained in:
blessedcoolant
2023-08-27 02:54:19 +12:00
parent c49851e027 61224e5cfe
commit e9633a3adb
388 changed files with 11955 additions and 10525 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
invokeai/app/api/routers/app_info.py
View File

@ -55,7 +55,7 @@ async def get_version() -> AppVersion:
@app_router.get("/config", operation_id="get_config", status_code=200, response_model=AppConfig)
async def get_config() -> AppConfig:
infill_methods = ["tile"]
infill_methods = ["tile", "lama"]
if PatchMatch.patchmatch_available():
infill_methods.append("patchmatch")

29
invokeai/app/api_app.py
View File

@ -1,11 +1,11 @@
# Copyright (c) 2022-2023 Kyle Schouviller (https://github.com/kyle0654) and the InvokeAI Team
import asyncio
from inspect import signature
import logging
import uvicorn
import socket
from inspect import signature
from pathlib import Path
import uvicorn
from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from fastapi.openapi.docs import get_redoc_html, get_swagger_ui_html
@ -13,7 +13,6 @@ from fastapi.openapi.utils import get_openapi
from fastapi.staticfiles import StaticFiles
from fastapi_events.handlers.local import local_handler
from fastapi_events.middleware import EventHandlerASGIMiddleware
from pathlib import Path
from pydantic.schema import schema
from .services.config import InvokeAIAppConfig
@ -30,9 +29,12 @@ from .api.sockets import SocketIO
from .invocations.baseinvocation import BaseInvocation, _InputField, _OutputField, UIConfigBase
import torch
# noinspection PyUnresolvedReferences
import invokeai.backend.util.hotfixes # noqa: F401 (monkeypatching on import)
if torch.backends.mps.is_available():
# noinspection PyUnresolvedReferences
import invokeai.backend.util.mps_fixes # noqa: F401 (monkeypatching on import)
@ -40,7 +42,6 @@ app_config = InvokeAIAppConfig.get_config()
app_config.parse_args()
logger = InvokeAILogger.getLogger(config=app_config)
# fix for windows mimetypes registry entries being borked
# see https://github.com/invoke-ai/InvokeAI/discussions/3684#discussioncomment-6391352
mimetypes.add_type("application/javascript", ".js")
@ -122,6 +123,7 @@ def custom_openapi():
output_schemas = schema(output_types, ref_prefix="#/components/schemas/")
for schema_key, output_schema in output_schemas["definitions"].items():
output_schema["class"] = "output"
openapi_schema["components"]["schemas"][schema_key] = output_schema
# TODO: note that we assume the schema_key here is the TYPE.__name__
@ -130,8 +132,8 @@ def custom_openapi():
# Add Node Editor UI helper schemas
ui_config_schemas = schema([UIConfigBase, _InputField, _OutputField], ref_prefix="#/components/schemas/")
for schema_key, output_schema in ui_config_schemas["definitions"].items():
openapi_schema["components"]["schemas"][schema_key] = output_schema
for schema_key, ui_config_schema in ui_config_schemas["definitions"].items():
openapi_schema["components"]["schemas"][schema_key] = ui_config_schema
# Add a reference to the output type to additionalProperties of the invoker schema
for invoker in all_invocations:
@ -140,8 +142,8 @@ def custom_openapi():
output_type_title = output_type_titles[output_type.__name__]
invoker_schema = openapi_schema["components"]["schemas"][invoker_name]
outputs_ref = {"$ref": f"#/components/schemas/{output_type_title}"}
invoker_schema["output"] = outputs_ref
invoker_schema["class"] = "invocation"
from invokeai.backend.model_management.models import get_model_config_enums
@ -207,6 +209,17 @@ def invoke_api():
check_invokeai_root(app_config) # note, may exit with an exception if root not set up
if app_config.dev_reload:
try:
import jurigged
except ImportError as e:
logger.error(
'Can\'t start `--dev_reload` because jurigged is not found; `pip install -e ".[dev]"` to include development dependencies.',
exc_info=e,
)
else:
jurigged.watch(logger=InvokeAILogger.getLogger(name="jurigged").info)
port = find_port(app_config.port)
if port != app_config.port:
logger.warn(f"Port {app_config.port} in use, using port {port}")

26
invokeai/app/invocations/baseinvocation.py
View File

@ -71,6 +71,9 @@ class FieldDescriptions:
safe_mode = "Whether or not to use safe mode"
scribble_mode = "Whether or not to use scribble mode"
scale_factor = "The factor by which to scale"
blend_alpha = (
"Blending factor. 0.0 = use input A only, 1.0 = use input B only, 0.5 = 50% mix of input A and input B."
)
num_1 = "The first number"
num_2 = "The second number"
mask = "The mask to use for the operation"
@ -140,6 +143,7 @@ class UIType(str, Enum):
# region Misc
FilePath = "FilePath"
Enum = "enum"
Scheduler = "Scheduler"
# endregion
@ -166,6 +170,7 @@ class _InputField(BaseModel):
ui_hidden: bool
ui_type: Optional[UIType]
ui_component: Optional[UIComponent]
ui_order: Optional[int]
class _OutputField(BaseModel):
@ -178,6 +183,7 @@ class _OutputField(BaseModel):
ui_hidden: bool
ui_type: Optional[UIType]
ui_order: Optional[int]
def InputField(
@ -211,6 +217,7 @@ def InputField(
ui_type: Optional[UIType] = None,
ui_component: Optional[UIComponent] = None,
ui_hidden: bool = False,
ui_order: Optional[int] = None,
**kwargs: Any,
) -> Any:
"""
@ -269,6 +276,7 @@ def InputField(
ui_type=ui_type,
ui_component=ui_component,
ui_hidden=ui_hidden,
ui_order=ui_order,
**kwargs,
)
@ -302,6 +310,7 @@ def OutputField(
repr: bool = True,
ui_type: Optional[UIType] = None,
ui_hidden: bool = False,
ui_order: Optional[int] = None,
**kwargs: Any,
) -> Any:
"""
@ -348,6 +357,7 @@ def OutputField(
repr=repr,
ui_type=ui_type,
ui_hidden=ui_hidden,
ui_order=ui_order,
**kwargs,
)
@ -376,7 +386,7 @@ class BaseInvocationOutput(BaseModel):
"""Base class for all invocation outputs"""
# All outputs must include a type name like this:
# type: Literal['your_output_name']
# type: Literal['your_output_name'] # noqa f821
@classmethod
def get_all_subclasses_tuple(cls):
@ -389,6 +399,13 @@ class BaseInvocationOutput(BaseModel):
toprocess.extend(next_subclasses)
return tuple(subclasses)
class Config:
@staticmethod
def schema_extra(schema: dict[str, Any], model_class: Type[BaseModel]) -> None:
if "required" not in schema or not isinstance(schema["required"], list):
schema["required"] = list()
schema["required"].extend(["type"])
class RequiredConnectionException(Exception):
"""Raised when an field which requires a connection did not receive a value."""
@ -410,7 +427,7 @@ class BaseInvocation(ABC, BaseModel):
"""
# All invocations must include a type name like this:
# type: Literal['your_output_name']
# type: Literal['your_output_name'] # noqa f821
@classmethod
def get_all_subclasses(cls):
@ -449,6 +466,9 @@ class BaseInvocation(ABC, BaseModel):
schema["title"] = uiconfig.title
if uiconfig and hasattr(uiconfig, "tags"):
schema["tags"] = uiconfig.tags
if "required" not in schema or not isinstance(schema["required"], list):
schema["required"] = list()
schema["required"].extend(["type", "id"])
@abstractmethod
def invoke(self, context: InvocationContext) -> BaseInvocationOutput:
@ -485,7 +505,7 @@ class BaseInvocation(ABC, BaseModel):
raise MissingInputException(self.__fields__["type"].default, field_name)
return self.invoke(context)
id: str = InputField(description="The id of this node. Must be unique among all nodes.")
id: str = Field(description="The id of this node. Must be unique among all nodes.")
is_intermediate: bool = InputField(
default=False, description="Whether or not this node is an intermediate node.", input=Input.Direct
)

2
invokeai/app/invocations/compel.py
View File

@ -233,7 +233,7 @@ class SDXLPromptInvocationBase:
dtype_for_device_getter=torch_dtype,
truncate_long_prompts=True, # TODO:
returned_embeddings_type=ReturnedEmbeddingsType.PENULTIMATE_HIDDEN_STATES_NON_NORMALIZED, # TODO: clip skip
requires_pooled=True,
requires_pooled=get_pooled,
)
conjunction = Compel.parse_prompt_string(prompt)

35
invokeai/app/invocations/image.py
View File

@ -8,7 +8,7 @@ import numpy
from PIL import Image, ImageChops, ImageFilter, ImageOps
from invokeai.app.invocations.metadata import CoreMetadata
from invokeai.app.invocations.primitives import ImageField, ImageOutput
from invokeai.app.invocations.primitives import ColorField, ImageField, ImageOutput
from invokeai.backend.image_util.invisible_watermark import InvisibleWatermark
from invokeai.backend.image_util.safety_checker import SafetyChecker
@ -41,6 +41,39 @@ class ShowImageInvocation(BaseInvocation):
)
@title("Blank Image")
@tags("image")
class BlankImageInvocation(BaseInvocation):
"""Creates a blank image and forwards it to the pipeline"""
# Metadata
type: Literal["blank_image"] = "blank_image"
# Inputs
width: int = InputField(default=512, description="The width of the image")
height: int = InputField(default=512, description="The height of the image")
mode: Literal["RGB", "RGBA"] = InputField(default="RGB", description="The mode of the image")
color: ColorField = InputField(default=ColorField(r=0, g=0, b=0, a=255), description="The color of the image")
def invoke(self, context: InvocationContext) -> ImageOutput:
image = Image.new(mode=self.mode, size=(self.width, self.height), color=self.color.tuple())
image_dto = context.services.images.create(
image=image,
image_origin=ResourceOrigin.INTERNAL,
image_category=ImageCategory.GENERAL,
node_id=self.id,
session_id=context.graph_execution_state_id,
is_intermediate=self.is_intermediate,
)
return ImageOutput(
image=ImageField(image_name=image_dto.image_name),
width=image_dto.width,
height=image_dto.height,
)
@title("Crop Image")
@tags("image", "crop")
class ImageCropInvocation(BaseInvocation):

46
invokeai/app/invocations/infill.py
View File

@ -1,23 +1,25 @@
# Copyright (c) 2022 Kyle Schouviller (https://github.com/kyle0654) and the InvokeAI Team
import math
from typing import Literal, Optional, get_args
import numpy as np
import math
from PIL import Image, ImageOps
from invokeai.app.invocations.primitives import ImageField, ImageOutput, ColorField
from invokeai.app.invocations.primitives import ColorField, ImageField, ImageOutput
from invokeai.app.util.misc import SEED_MAX, get_random_seed
from invokeai.backend.image_util.lama import LaMA
from invokeai.backend.image_util.patchmatch import PatchMatch
from ..models.image import ImageCategory, ResourceOrigin
from .baseinvocation import BaseInvocation, InputField, InvocationContext, title, tags
from .baseinvocation import BaseInvocation, InputField, InvocationContext, tags, title
def infill_methods() -> list[str]:
methods = [
"tile",
"solid",
"lama",
]
if PatchMatch.patchmatch_available():
methods.insert(0, "patchmatch")
@ -28,6 +30,11 @@ INFILL_METHODS = Literal[tuple(infill_methods())]
DEFAULT_INFILL_METHOD = "patchmatch" if "patchmatch" in get_args(INFILL_METHODS) else "tile"
def infill_lama(im: Image.Image) -> Image.Image:
lama = LaMA()
return lama(im)
def infill_patchmatch(im: Image.Image) -> Image.Image:
if im.mode != "RGBA":
return im
@ -90,7 +97,7 @@ def tile_fill_missing(im: Image.Image, tile_size: int = 16, seed: Optional[int]
return im
# Find all invalid tiles and replace with a random valid tile
replace_count = (tiles_mask is False).sum()
replace_count = (tiles_mask == False).sum() # noqa: E712
rng = np.random.default_rng(seed=seed)
tiles_all[np.logical_not(tiles_mask)] = filtered_tiles[rng.choice(filtered_tiles.shape[0], replace_count), :, :, :]
@ -218,3 +225,34 @@ class InfillPatchMatchInvocation(BaseInvocation):
width=image_dto.width,
height=image_dto.height,
)
@title("LaMa Infill")
@tags("image", "inpaint")
class LaMaInfillInvocation(BaseInvocation):
"""Infills transparent areas of an image using the LaMa model"""
type: Literal["infill_lama"] = "infill_lama"
# Inputs
image: ImageField = InputField(description="The image to infill")
def invoke(self, context: InvocationContext) -> ImageOutput:
image = context.services.images.get_pil_image(self.image.image_name)
infilled = infill_lama(image.copy())
image_dto = context.services.images.create(
image=infilled,
image_origin=ResourceOrigin.INTERNAL,
image_category=ImageCategory.GENERAL,
node_id=self.id,
session_id=context.graph_execution_state_id,
is_intermediate=self.is_intermediate,
)
return ImageOutput(
image=ImageField(image_name=image_dto.image_name),
width=image_dto.width,
height=image_dto.height,
)

99
invokeai/app/invocations/latent.py
View File

@ -4,6 +4,7 @@ from contextlib import ExitStack
from typing import List, Literal, Optional, Union
import einops
import numpy as np
import torch
import torchvision.transforms as T
from diffusers.image_processor import VaeImageProcessor
@ -168,22 +169,24 @@ class DenoiseLatentsInvocation(BaseInvocation):
# Inputs
positive_conditioning: ConditioningField = InputField(
description=FieldDescriptions.positive_cond, input=Input.Connection
description=FieldDescriptions.positive_cond, input=Input.Connection, ui_order=0
)
negative_conditioning: ConditioningField = InputField(
description=FieldDescriptions.negative_cond, input=Input.Connection
description=FieldDescriptions.negative_cond, input=Input.Connection, ui_order=1
)
noise: Optional[LatentsField] = InputField(description=FieldDescriptions.noise, input=Input.Connection)
noise: Optional[LatentsField] = InputField(description=FieldDescriptions.noise, input=Input.Connection, ui_order=3)
steps: int = InputField(default=10, gt=0, description=FieldDescriptions.steps)
cfg_scale: Union[float, List[float]] = InputField(
default=7.5, ge=1, description=FieldDescriptions.cfg_scale, ui_type=UIType.Float
default=7.5, ge=1, description=FieldDescriptions.cfg_scale, ui_type=UIType.Float, title="CFG Scale"
)
denoising_start: float = InputField(default=0.0, ge=0, le=1, description=FieldDescriptions.denoising_start)
denoising_end: float = InputField(default=1.0, ge=0, le=1, description=FieldDescriptions.denoising_end)
scheduler: SAMPLER_NAME_VALUES = InputField(default="euler", description=FieldDescriptions.scheduler)
unet: UNetField = InputField(description=FieldDescriptions.unet, input=Input.Connection)
scheduler: SAMPLER_NAME_VALUES = InputField(
default="euler", description=FieldDescriptions.scheduler, ui_type=UIType.Scheduler
)
unet: UNetField = InputField(description=FieldDescriptions.unet, input=Input.Connection, title="UNet", ui_order=2)
control: Union[ControlField, list[ControlField]] = InputField(
default=None, description=FieldDescriptions.control, input=Input.Connection
default=None, description=FieldDescriptions.control, input=Input.Connection, ui_order=5
)
latents: Optional[LatentsField] = InputField(description=FieldDescriptions.latents, input=Input.Connection)
mask: Optional[InpaintMaskField] = InputField(
@ -517,7 +520,7 @@ class DenoiseLatentsInvocation(BaseInvocation):
@title("Latents to Image")
@tags("latents", "image", "vae")
@tags("latents", "image", "vae", "l2i")
class LatentsToImageInvocation(BaseInvocation):
"""Generates an image from latents."""
@ -705,7 +708,7 @@ class ScaleLatentsInvocation(BaseInvocation):
@title("Image to Latents")
@tags("latents", "image", "vae")
@tags("latents", "image", "vae", "i2l")
class ImageToLatentsInvocation(BaseInvocation):
"""Encodes an image into latents."""
@ -786,3 +789,81 @@ class ImageToLatentsInvocation(BaseInvocation):
latents = latents.to("cpu")
context.services.latents.save(name, latents)
return build_latents_output(latents_name=name, latents=latents, seed=None)
@title("Blend Latents")
@tags("latents", "blend")
class BlendLatentsInvocation(BaseInvocation):
"""Blend two latents using a given alpha. Latents must have same size."""
type: Literal["lblend"] = "lblend"
# Inputs
latents_a: LatentsField = InputField(
description=FieldDescriptions.latents,
input=Input.Connection,
)
latents_b: LatentsField = InputField(
description=FieldDescriptions.latents,
input=Input.Connection,
)
alpha: float = InputField(default=0.5, description=FieldDescriptions.blend_alpha)
def invoke(self, context: InvocationContext) -> LatentsOutput:
latents_a = context.services.latents.get(self.latents_a.latents_name)
latents_b = context.services.latents.get(self.latents_b.latents_name)
if latents_a.shape != latents_b.shape:
raise "Latents to blend must be the same size."
# TODO:
device = choose_torch_device()
def slerp(t, v0, v1, DOT_THRESHOLD=0.9995):
"""
Spherical linear interpolation
Args:
t (float/np.ndarray): Float value between 0.0 and 1.0
v0 (np.ndarray): Starting vector
v1 (np.ndarray): Final vector
DOT_THRESHOLD (float): Threshold for considering the two vectors as
colineal. Not recommended to alter this.
Returns:
v2 (np.ndarray): Interpolation vector between v0 and v1
"""
inputs_are_torch = False
if not isinstance(v0, np.ndarray):
inputs_are_torch = True
v0 = v0.detach().cpu().numpy()
if not isinstance(v1, np.ndarray):
inputs_are_torch = True
v1 = v1.detach().cpu().numpy()
dot = np.sum(v0 * v1 / (np.linalg.norm(v0) * np.linalg.norm(v1)))
if np.abs(dot) > DOT_THRESHOLD:
v2 = (1 - t) * v0 + t * v1
else:
theta_0 = np.arccos(dot)
sin_theta_0 = np.sin(theta_0)
theta_t = theta_0 * t
sin_theta_t = np.sin(theta_t)
s0 = np.sin(theta_0 - theta_t) / sin_theta_0
s1 = sin_theta_t / sin_theta_0
v2 = s0 * v0 + s1 * v1
if inputs_are_torch:
v2 = torch.from_numpy(v2).to(device)
return v2
# blend
blended_latents = slerp(self.alpha, latents_a, latents_b)
# https://discuss.huggingface.co/t/memory-usage-by-later-pipeline-stages/23699
blended_latents = blended_latents.to("cpu")
torch.cuda.empty_cache()
name = f"{context.graph_execution_state_id}__{self.id}"
# context.services.latents.set(name, resized_latents)
context.services.latents.save(name, blended_latents)
return build_latents_output(latents_name=name, latents=blended_latents)

10
invokeai/app/invocations/math.py
View File

@ -21,7 +21,7 @@ class AddInvocation(BaseInvocation):
b: int = InputField(default=0, description=FieldDescriptions.num_2)
def invoke(self, context: InvocationContext) -> IntegerOutput:
return IntegerOutput(a=self.a + self.b)
return IntegerOutput(value=self.a + self.b)
@title("Subtract Integers")
@ -36,7 +36,7 @@ class SubtractInvocation(BaseInvocation):
b: int = InputField(default=0, description=FieldDescriptions.num_2)
def invoke(self, context: InvocationContext) -> IntegerOutput:
return IntegerOutput(a=self.a - self.b)
return IntegerOutput(value=self.a - self.b)
@title("Multiply Integers")
@ -51,7 +51,7 @@ class MultiplyInvocation(BaseInvocation):
b: int = InputField(default=0, description=FieldDescriptions.num_2)
def invoke(self, context: InvocationContext) -> IntegerOutput:
return IntegerOutput(a=self.a * self.b)
return IntegerOutput(value=self.a * self.b)
@title("Divide Integers")
@ -66,7 +66,7 @@ class DivideInvocation(BaseInvocation):
b: int = InputField(default=0, description=FieldDescriptions.num_2)
def invoke(self, context: InvocationContext) -> IntegerOutput:
return IntegerOutput(a=int(self.a / self.b))
return IntegerOutput(value=int(self.a / self.b))
@title("Random Integer")
@ -81,4 +81,4 @@ class RandomIntInvocation(BaseInvocation):
high: int = InputField(default=np.iinfo(np.int32).max, description="The exclusive high value")
def invoke(self, context: InvocationContext) -> IntegerOutput:
return IntegerOutput(a=np.random.randint(self.low, self.high))
return IntegerOutput(value=np.random.randint(self.low, self.high))

1
invokeai/app/invocations/metadata.py
View File

@ -32,6 +32,7 @@ class CoreMetadata(BaseModelExcludeNull):
generation_mode: str = Field(
description="The generation mode that output this image",
)
created_by: Optional[str] = Field(description="The name of the creator of the image")
positive_prompt: str = Field(description="The positive prompt parameter")
negative_prompt: str = Field(description="The negative prompt parameter")
width: int = Field(description="The width parameter")

8
invokeai/app/invocations/model.py
View File

@ -72,7 +72,7 @@ class LoRAModelField(BaseModel):
base_model: BaseModelType = Field(description="Base model")
@title("Main Model Loader")
@title("Main Model")
@tags("model")
class MainModelLoaderInvocation(BaseInvocation):
"""Loads a main model, outputting its submodels."""
@ -179,7 +179,7 @@ class LoraLoaderOutput(BaseInvocationOutput):
# fmt: on
@title("LoRA Loader")
@title("LoRA")
@tags("lora", "model")
class LoraLoaderInvocation(BaseInvocation):
"""Apply selected lora to unet and text_encoder."""
@ -257,7 +257,7 @@ class SDXLLoraLoaderOutput(BaseInvocationOutput):
# fmt: on
@title("SDXL LoRA Loader")
@title("SDXL LoRA")
@tags("sdxl", "lora", "model")
class SDXLLoraLoaderInvocation(BaseInvocation):
"""Apply selected lora to unet and text_encoder."""
@ -356,7 +356,7 @@ class VaeLoaderOutput(BaseInvocationOutput):
vae: VaeField = OutputField(description=FieldDescriptions.vae, title="VAE")
@title("VAE Loader")
@title("VAE")
@tags("vae", "model")
class VaeLoaderInvocation(BaseInvocation):
"""Loads a VAE model, outputting a VaeLoaderOutput"""

4
invokeai/app/invocations/onnx.py
View File

@ -169,7 +169,7 @@ class ONNXTextToLatentsInvocation(BaseInvocation):
ui_type=UIType.Float,
)
scheduler: SAMPLER_NAME_VALUES = InputField(
default="euler", description=FieldDescriptions.scheduler, input=Input.Direct
default="euler", description=FieldDescriptions.scheduler, input=Input.Direct, ui_type=UIType.Scheduler
)
precision: PRECISION_VALUES = InputField(default="tensor(float16)", description=FieldDescriptions.precision)
unet: UNetField = InputField(
@ -406,7 +406,7 @@ class OnnxModelField(BaseModel):
model_type: ModelType = Field(description="Model Type")
@title("ONNX Model Loader")
@title("ONNX Main Model")
@tags("onnx", "model")
class OnnxModelLoaderInvocation(BaseInvocation):
"""Loads a main model, outputting its submodels."""

58
invokeai/app/invocations/primitives.py
View File

@ -33,7 +33,7 @@ class BooleanOutput(BaseInvocationOutput):
"""Base class for nodes that output a single boolean"""
type: Literal["boolean_output"] = "boolean_output"
a: bool = OutputField(description="The output boolean")
value: bool = OutputField(description="The output boolean")
class BooleanCollectionOutput(BaseInvocationOutput):
@ -42,9 +42,7 @@ class BooleanCollectionOutput(BaseInvocationOutput):
type: Literal["boolean_collection_output"] = "boolean_collection_output"
# Outputs
collection: list[bool] = OutputField(
default_factory=list, description="The output boolean collection", ui_type=UIType.BooleanCollection
)
collection: list[bool] = OutputField(description="The output boolean collection", ui_type=UIType.BooleanCollection)
@title("Boolean Primitive")
@ -55,10 +53,10 @@ class BooleanInvocation(BaseInvocation):
type: Literal["boolean"] = "boolean"
# Inputs
a: bool = InputField(default=False, description="The boolean value")
value: bool = InputField(default=False, description="The boolean value")
def invoke(self, context: InvocationContext) -> BooleanOutput:
return BooleanOutput(a=self.a)
return BooleanOutput(value=self.value)
@title("Boolean Primitive Collection")
@ -70,7 +68,7 @@ class BooleanCollectionInvocation(BaseInvocation):
# Inputs
collection: list[bool] = InputField(
default=False, description="The collection of boolean values", ui_type=UIType.BooleanCollection
default_factory=list, description="The collection of boolean values", ui_type=UIType.BooleanCollection
)
def invoke(self, context: InvocationContext) -> BooleanCollectionOutput:
@ -86,7 +84,7 @@ class IntegerOutput(BaseInvocationOutput):
"""Base class for nodes that output a single integer"""
type: Literal["integer_output"] = "integer_output"
a: int = OutputField(description="The output integer")
value: int = OutputField(description="The output integer")
class IntegerCollectionOutput(BaseInvocationOutput):
@ -95,9 +93,7 @@ class IntegerCollectionOutput(BaseInvocationOutput):
type: Literal["integer_collection_output"] = "integer_collection_output"
# Outputs
collection: list[int] = OutputField(
default_factory=list, description="The int collection", ui_type=UIType.IntegerCollection
)
collection: list[int] = OutputField(description="The int collection", ui_type=UIType.IntegerCollection)
@title("Integer Primitive")
@ -108,10 +104,10 @@ class IntegerInvocation(BaseInvocation):
type: Literal["integer"] = "integer"
# Inputs
a: int = InputField(default=0, description="The integer value")
value: int = InputField(default=0, description="The integer value")
def invoke(self, context: InvocationContext) -> IntegerOutput:
return IntegerOutput(a=self.a)
return IntegerOutput(value=self.value)
@title("Integer Primitive Collection")
@ -139,7 +135,7 @@ class FloatOutput(BaseInvocationOutput):
"""Base class for nodes that output a single float"""
type: Literal["float_output"] = "float_output"
a: float = OutputField(description="The output float")
value: float = OutputField(description="The output float")
class FloatCollectionOutput(BaseInvocationOutput):
@ -148,9 +144,7 @@ class FloatCollectionOutput(BaseInvocationOutput):
type: Literal["float_collection_output"] = "float_collection_output"
# Outputs
collection: list[float] = OutputField(
default_factory=list, description="The float collection", ui_type=UIType.FloatCollection
)
collection: list[float] = OutputField(description="The float collection", ui_type=UIType.FloatCollection)
@title("Float Primitive")
@ -161,10 +155,10 @@ class FloatInvocation(BaseInvocation):
type: Literal["float"] = "float"
# Inputs
param: float = InputField(default=0.0, description="The float value")
value: float = InputField(default=0.0, description="The float value")
def invoke(self, context: InvocationContext) -> FloatOutput:
return FloatOutput(a=self.param)
return FloatOutput(value=self.value)
@title("Float Primitive Collection")
@ -176,7 +170,7 @@ class FloatCollectionInvocation(BaseInvocation):
# Inputs
collection: list[float] = InputField(
default=0, description="The collection of float values", ui_type=UIType.FloatCollection
default_factory=list, description="The collection of float values", ui_type=UIType.FloatCollection
)
def invoke(self, context: InvocationContext) -> FloatCollectionOutput:
@ -192,7 +186,7 @@ class StringOutput(BaseInvocationOutput):
"""Base class for nodes that output a single string"""
type: Literal["string_output"] = "string_output"
text: str = OutputField(description="The output string")
value: str = OutputField(description="The output string")
class StringCollectionOutput(BaseInvocationOutput):
@ -201,9 +195,7 @@ class StringCollectionOutput(BaseInvocationOutput):
type: Literal["string_collection_output"] = "string_collection_output"
# Outputs
collection: list[str] = OutputField(
default_factory=list, description="The output strings", ui_type=UIType.StringCollection
)
collection: list[str] = OutputField(description="The output strings", ui_type=UIType.StringCollection)
@title("String Primitive")
@ -214,10 +206,10 @@ class StringInvocation(BaseInvocation):
type: Literal["string"] = "string"
# Inputs
text: str = InputField(default="", description="The string value", ui_component=UIComponent.Textarea)
value: str = InputField(default="", description="The string value", ui_component=UIComponent.Textarea)
def invoke(self, context: InvocationContext) -> StringOutput:
return StringOutput(text=self.text)
return StringOutput(value=self.value)
@title("String Primitive Collection")
@ -229,7 +221,7 @@ class StringCollectionInvocation(BaseInvocation):
# Inputs
collection: list[str] = InputField(
default=0, description="The collection of string values", ui_type=UIType.StringCollection
default_factory=list, description="The collection of string values", ui_type=UIType.StringCollection
)
def invoke(self, context: InvocationContext) -> StringCollectionOutput:
@ -262,9 +254,7 @@ class ImageCollectionOutput(BaseInvocationOutput):
type: Literal["image_collection_output"] = "image_collection_output"
# Outputs
collection: list[ImageField] = OutputField(
default_factory=list, description="The output images", ui_type=UIType.ImageCollection
)
collection: list[ImageField] = OutputField(description="The output images", ui_type=UIType.ImageCollection)
@title("Image Primitive")
@ -353,7 +343,6 @@ class LatentsCollectionOutput(BaseInvocationOutput):
type: Literal["latents_collection_output"] = "latents_collection_output"
collection: list[LatentsField] = OutputField(
default_factory=list,
description=FieldDescriptions.latents,
ui_type=UIType.LatentsCollection,
)
@ -384,7 +373,7 @@ class LatentsCollectionInvocation(BaseInvocation):
# Inputs
collection: list[LatentsField] = InputField(
default=0, description="The collection of latents tensors", ui_type=UIType.LatentsCollection
description="The collection of latents tensors", ui_type=UIType.LatentsCollection
)
def invoke(self, context: InvocationContext) -> LatentsCollectionOutput:
@ -429,9 +418,7 @@ class ColorCollectionOutput(BaseInvocationOutput):
type: Literal["color_collection_output"] = "color_collection_output"
# Outputs
collection: list[ColorField] = OutputField(
default_factory=list, description="The output colors", ui_type=UIType.ColorCollection
)
collection: list[ColorField] = OutputField(description="The output colors", ui_type=UIType.ColorCollection)
@title("Color Primitive")
@ -474,7 +461,6 @@ class ConditioningCollectionOutput(BaseInvocationOutput):
# Outputs
collection: list[ConditioningField] = OutputField(
default_factory=list,
description="The output conditioning tensors",
ui_type=UIType.ConditioningCollection,
)

4
invokeai/app/invocations/sdxl.py
View File

@ -37,7 +37,7 @@ class SDXLRefinerModelLoaderOutput(BaseInvocationOutput):
vae: VaeField = OutputField(description=FieldDescriptions.vae, title="VAE")
@title("SDXL Main Model Loader")
@title("SDXL Main Model")
@tags("model", "sdxl")
class SDXLModelLoaderInvocation(BaseInvocation):
"""Loads an sdxl base model, outputting its submodels."""
@ -122,7 +122,7 @@ class SDXLModelLoaderInvocation(BaseInvocation):
)
@title("SDXL Refiner Model Loader")
@title("SDXL Refiner Model")
@tags("model", "sdxl", "refiner")
class SDXLRefinerModelLoaderInvocation(BaseInvocation):
"""Loads an sdxl refiner model, outputting its submodels."""

8
invokeai/app/services/config/__init__.py Normal file
View File

@ -0,0 +1,8 @@
"""
Init file for InvokeAI configure package
"""
from .invokeai_config import ( # noqa F401
InvokeAIAppConfig,
get_invokeai_config,
)

239
invokeai/app/services/config/base.py Normal file
View File

@ -0,0 +1,239 @@
# Copyright (c) 2023 Lincoln Stein (https://github.com/lstein) and the InvokeAI Development Team
"""
Base class for the InvokeAI configuration system.
It defines a type of pydantic BaseSettings object that
is able to read and write from an omegaconf-based config file,
with overriding of settings from environment variables and/or
the command line.
"""
from __future__ import annotations
import argparse
import os
import pydoc
import sys
from argparse import ArgumentParser
from omegaconf import OmegaConf, DictConfig, ListConfig
from pathlib import Path
from pydantic import BaseSettings
from typing import ClassVar, Dict, List, Literal, Union, get_origin, get_type_hints, get_args
class PagingArgumentParser(argparse.ArgumentParser):
"""
A custom ArgumentParser that uses pydoc to page its output.
It also supports reading defaults from an init file.
"""
def print_help(self, file=None):
text = self.format_help()
pydoc.pager(text)
class InvokeAISettings(BaseSettings):
"""
Runtime configuration settings in which default values are
read from an omegaconf .yaml file.
"""
initconf: ClassVar[DictConfig] = None
argparse_groups: ClassVar[Dict] = {}
def parse_args(self, argv: list = sys.argv[1:]):
parser = self.get_parser()
opt = parser.parse_args(argv)
for name in self.__fields__:
if name not in self._excluded():
value = getattr(opt, name)
if isinstance(value, ListConfig):
value = list(value)
elif isinstance(value, DictConfig):
value = dict(value)
setattr(self, name, value)
def to_yaml(self) -> str:
"""
Return a YAML string representing our settings. This can be used
as the contents of `invokeai.yaml` to restore settings later.
"""
cls = self.__class__
type = get_args(get_type_hints(cls)["type"])[0]
field_dict = dict({type: dict()})
for name, field in self.__fields__.items():
if name in cls._excluded_from_yaml():
continue
category = field.field_info.extra.get("category") or "Uncategorized"
value = getattr(self, name)
if category not in field_dict[type]:
field_dict[type][category] = dict()
# keep paths as strings to make it easier to read
field_dict[type][category][name] = str(value) if isinstance(value, Path) else value
conf = OmegaConf.create(field_dict)
return OmegaConf.to_yaml(conf)
@classmethod
def add_parser_arguments(cls, parser):
if "type" in get_type_hints(cls):
settings_stanza = get_args(get_type_hints(cls)["type"])[0]
else:
settings_stanza = "Uncategorized"
env_prefix = cls.Config.env_prefix if hasattr(cls.Config, "env_prefix") else settings_stanza.upper()
initconf = (
cls.initconf.get(settings_stanza)
if cls.initconf and settings_stanza in cls.initconf
else OmegaConf.create()
)
# create an upcase version of the environment in
# order to achieve case-insensitive environment
# variables (the way Windows does)
upcase_environ = dict()
for key, value in os.environ.items():
upcase_environ[key.upper()] = value
fields = cls.__fields__
cls.argparse_groups = {}
for name, field in fields.items():
if name not in cls._excluded():
current_default = field.default
category = field.field_info.extra.get("category", "Uncategorized")
env_name = env_prefix + "_" + name
if category in initconf and name in initconf.get(category):
field.default = initconf.get(category).get(name)
if env_name.upper() in upcase_environ:
field.default = upcase_environ[env_name.upper()]
cls.add_field_argument(parser, name, field)
field.default = current_default
@classmethod
def cmd_name(self, command_field: str = "type") -> str:
hints = get_type_hints(self)
if command_field in hints:
return get_args(hints[command_field])[0]
else:
return "Uncategorized"
@classmethod
def get_parser(cls) -> ArgumentParser:
parser = PagingArgumentParser(
prog=cls.cmd_name(),
description=cls.__doc__,
)
cls.add_parser_arguments(parser)
return parser
@classmethod
def add_subparser(cls, parser: argparse.ArgumentParser):
parser.add_parser(cls.cmd_name(), help=cls.__doc__)
@classmethod
def _excluded(self) -> List[str]:
# internal fields that shouldn't be exposed as command line options
return ["type", "initconf"]
@classmethod
def _excluded_from_yaml(self) -> List[str]:
# combination of deprecated parameters and internal ones that shouldn't be exposed as invokeai.yaml options
return [
"type",
"initconf",
"version",
"from_file",
"model",
"root",
"max_cache_size",
"max_vram_cache_size",
"always_use_cpu",
"free_gpu_mem",
"xformers_enabled",
"tiled_decode",
]
class Config:
env_file_encoding = "utf-8"
arbitrary_types_allowed = True
case_sensitive = True
@classmethod
def add_field_argument(cls, command_parser, name: str, field, default_override=None):
field_type = get_type_hints(cls).get(name)
default = (
default_override
if default_override is not None
else field.default
if field.default_factory is None
else field.default_factory()
)
if category := field.field_info.extra.get("category"):
if category not in cls.argparse_groups:
cls.argparse_groups[category] = command_parser.add_argument_group(category)
argparse_group = cls.argparse_groups[category]
else:
argparse_group = command_parser
if get_origin(field_type) == Literal:
allowed_values = get_args(field.type_)
allowed_types = set()
for val in allowed_values:
allowed_types.add(type(val))
allowed_types_list = list(allowed_types)
field_type = allowed_types_list[0] if len(allowed_types) == 1 else int_or_float_or_str
argparse_group.add_argument(
f"--{name}",
dest=name,
type=field_type,
default=default,
choices=allowed_values,
help=field.field_info.description,
)
elif get_origin(field_type) == Union:
argparse_group.add_argument(
f"--{name}",
dest=name,
type=int_or_float_or_str,
default=default,
help=field.field_info.description,
)
elif get_origin(field_type) == list:
argparse_group.add_argument(
f"--{name}",
dest=name,
nargs="*",
type=field.type_,
default=default,
action=argparse.BooleanOptionalAction if field.type_ == bool else "store",
help=field.field_info.description,
)
else:
argparse_group.add_argument(
f"--{name}",
dest=name,
type=field.type_,
default=default,
action=argparse.BooleanOptionalAction if field.type_ == bool else "store",
help=field.field_info.description,
)
def int_or_float_or_str(value: str) -> Union[int, float, str]:
"""
Workaround for argparse type checking.
"""
try:
return int(value)
except Exception as e: # noqa F841
pass
try:
return float(value)
except Exception as e: # noqa F841
pass
return str(value)

374
invokeai/app/services/config.py → invokeai/app/services/config/invokeai_config.py
View File

@ -10,37 +10,49 @@ categories returned by `invokeai --help`. The file looks like this:
[file: invokeai.yaml]
InvokeAI:
Paths:
root: /home/lstein/invokeai-main
conf_path: configs/models.yaml
legacy_conf_dir: configs/stable-diffusion
outdir: outputs
autoimport_dir: null
Models:
model: stable-diffusion-1.5
embeddings: true
Memory/Performance:
xformers_enabled: false
sequential_guidance: false
precision: float16
max_cache_size: 6
max_vram_cache_size: 0.5
always_use_cpu: false
free_gpu_mem: false
Features:
esrgan: true
patchmatch: true
internet_available: true
log_tokenization: false
Web Server:
host: 127.0.0.1
port: 8081
port: 9090
allow_origins: []
allow_credentials: true
allow_methods:
- '*'
allow_headers:
- '*'
Features:
esrgan: true
internet_available: true
log_tokenization: false
patchmatch: true
ignore_missing_core_models: false
Paths:
autoimport_dir: autoimport
lora_dir: null
embedding_dir: null
controlnet_dir: null
conf_path: configs/models.yaml
models_dir: models
legacy_conf_dir: configs/stable-diffusion
db_dir: databases
outdir: /home/lstein/invokeai-main/outputs
use_memory_db: false
Logging:
log_handlers:
- console
log_format: plain
log_level: info
Model Cache:
ram: 13.5
vram: 0.25
lazy_offload: true
Device:
device: auto
precision: auto
Generation:
sequential_guidance: false
attention_type: xformers
attention_slice_size: auto
force_tiled_decode: false
The default name of the configuration file is `invokeai.yaml`, located
in INVOKEAI_ROOT. You can replace supersede this by providing any
@ -54,24 +66,23 @@ InvokeAIAppConfig.parse_args() will parse the contents of `sys.argv`
at initialization time. You may pass a list of strings in the optional
`argv` argument to use instead of the system argv:
conf.parse_args(argv=['--xformers_enabled'])
conf.parse_args(argv=['--log_tokenization'])
It is also possible to set a value at initialization time. However, if
you call parse_args() it may be overwritten.
conf = InvokeAIAppConfig(xformers_enabled=True)
conf.parse_args(argv=['--no-xformers'])
conf.xformers_enabled
conf = InvokeAIAppConfig(log_tokenization=True)
conf.parse_args(argv=['--no-log_tokenization'])
conf.log_tokenization
# False
To avoid this, use `get_config()` to retrieve the application-wide
configuration object. This will retain any properties set at object
creation time:
conf = InvokeAIAppConfig.get_config(xformers_enabled=True)
conf.parse_args(argv=['--no-xformers'])
conf.xformers_enabled
conf = InvokeAIAppConfig.get_config(log_tokenization=True)
conf.parse_args(argv=['--no-log_tokenization'])
conf.log_tokenization
# True
Any setting can be overwritten by setting an environment variable of
@ -93,7 +104,7 @@ Typical usage at the top level file:
# get global configuration and print its cache size
conf = InvokeAIAppConfig.get_config()
conf.parse_args()
print(conf.max_cache_size)
print(conf.ram_cache_size)
Typical usage in a backend module:
@ -101,8 +112,7 @@ Typical usage in a backend module:
# get global configuration and print its cache size value
conf = InvokeAIAppConfig.get_config()
print(conf.max_cache_size)
print(conf.ram_cache_size)
Computed properties:
@ -159,15 +169,15 @@ two configs are kept in separate sections of the config file:
"""
from __future__ import annotations
import argparse
import pydoc
import os
import sys
from argparse import ArgumentParser
from omegaconf import OmegaConf, DictConfig, ListConfig
from pathlib import Path
from pydantic import BaseSettings, Field, parse_obj_as
from typing import ClassVar, Dict, List, Literal, Union, get_origin, get_type_hints, get_args
from typing import ClassVar, Dict, List, Literal, Union, get_type_hints, Optional
from omegaconf import OmegaConf, DictConfig
from pydantic import Field, parse_obj_as
from .base import InvokeAISettings
INIT_FILE = Path("invokeai.yaml")
DB_FILE = Path("invokeai.db")
@ -175,195 +185,6 @@ LEGACY_INIT_FILE = Path("invokeai.init")
DEFAULT_MAX_VRAM = 0.5
class InvokeAISettings(BaseSettings):
"""
Runtime configuration settings in which default values are
read from an omegaconf .yaml file.
"""
initconf: ClassVar[DictConfig] = None
argparse_groups: ClassVar[Dict] = {}
def parse_args(self, argv: list = sys.argv[1:]):
parser = self.get_parser()
opt = parser.parse_args(argv)
for name in self.__fields__:
if name not in self._excluded():
value = getattr(opt, name)
if isinstance(value, ListConfig):
value = list(value)
elif isinstance(value, DictConfig):
value = dict(value)
setattr(self, name, value)
def to_yaml(self) -> str:
"""
Return a YAML string representing our settings. This can be used
as the contents of `invokeai.yaml` to restore settings later.
"""
cls = self.__class__
type = get_args(get_type_hints(cls)["type"])[0]
field_dict = dict({type: dict()})
for name, field in self.__fields__.items():
if name in cls._excluded_from_yaml():
continue
category = field.field_info.extra.get("category") or "Uncategorized"
value = getattr(self, name)
if category not in field_dict[type]:
field_dict[type][category] = dict()
# keep paths as strings to make it easier to read
field_dict[type][category][name] = str(value) if isinstance(value, Path) else value
conf = OmegaConf.create(field_dict)
return OmegaConf.to_yaml(conf)
@classmethod
def add_parser_arguments(cls, parser):
if "type" in get_type_hints(cls):
settings_stanza = get_args(get_type_hints(cls)["type"])[0]
else:
settings_stanza = "Uncategorized"
env_prefix = cls.Config.env_prefix if hasattr(cls.Config, "env_prefix") else settings_stanza.upper()
initconf = (
cls.initconf.get(settings_stanza)
if cls.initconf and settings_stanza in cls.initconf
else OmegaConf.create()
)
# create an upcase version of the environment in
# order to achieve case-insensitive environment
# variables (the way Windows does)
upcase_environ = dict()
for key, value in os.environ.items():
upcase_environ[key.upper()] = value
fields = cls.__fields__
cls.argparse_groups = {}
for name, field in fields.items():
if name not in cls._excluded():
current_default = field.default
category = field.field_info.extra.get("category", "Uncategorized")
env_name = env_prefix + "_" + name
if category in initconf and name in initconf.get(category):
field.default = initconf.get(category).get(name)
if env_name.upper() in upcase_environ:
field.default = upcase_environ[env_name.upper()]
cls.add_field_argument(parser, name, field)
field.default = current_default
@classmethod
def cmd_name(self, command_field: str = "type") -> str:
hints = get_type_hints(self)
if command_field in hints:
return get_args(hints[command_field])[0]
else:
return "Uncategorized"
@classmethod
def get_parser(cls) -> ArgumentParser:
parser = PagingArgumentParser(
prog=cls.cmd_name(),
description=cls.__doc__,
)
cls.add_parser_arguments(parser)
return parser
@classmethod
def add_subparser(cls, parser: argparse.ArgumentParser):
parser.add_parser(cls.cmd_name(), help=cls.__doc__)
@classmethod
def _excluded(self) -> List[str]:
# internal fields that shouldn't be exposed as command line options
return ["type", "initconf"]
@classmethod
def _excluded_from_yaml(self) -> List[str]:
# combination of deprecated parameters and internal ones that shouldn't be exposed as invokeai.yaml options
return [
"type",
"initconf",
"version",
"from_file",
"model",
"root",
]
class Config:
env_file_encoding = "utf-8"
arbitrary_types_allowed = True
case_sensitive = True
@classmethod
def add_field_argument(cls, command_parser, name: str, field, default_override=None):
field_type = get_type_hints(cls).get(name)
default = (
default_override
if default_override is not None
else field.default
if field.default_factory is None
else field.default_factory()
)
if category := field.field_info.extra.get("category"):
if category not in cls.argparse_groups:
cls.argparse_groups[category] = command_parser.add_argument_group(category)
argparse_group = cls.argparse_groups[category]
else:
argparse_group = command_parser
if get_origin(field_type) == Literal:
allowed_values = get_args(field.type_)
allowed_types = set()
for val in allowed_values:
allowed_types.add(type(val))
allowed_types_list = list(allowed_types)
field_type = allowed_types_list[0] if len(allowed_types) == 1 else Union[allowed_types_list] # type: ignore
argparse_group.add_argument(
f"--{name}",
dest=name,
type=field_type,
default=default,
choices=allowed_values,
help=field.field_info.description,
)
elif get_origin(field_type) == list:
argparse_group.add_argument(
f"--{name}",
dest=name,
nargs="*",
type=field.type_,
default=default,
action=argparse.BooleanOptionalAction if field.type_ == bool else "store",
help=field.field_info.description,
)
else:
argparse_group.add_argument(
f"--{name}",
dest=name,
type=field.type_,
default=default,
action=argparse.BooleanOptionalAction if field.type_ == bool else "store",
help=field.field_info.description,
)
def _find_root() -> Path:
venv = Path(os.environ.get("VIRTUAL_ENV") or ".")
if os.environ.get("INVOKEAI_ROOT"):
root = Path(os.environ["INVOKEAI_ROOT"])
elif any([(venv.parent / x).exists() for x in [INIT_FILE, LEGACY_INIT_FILE]]):
root = (venv.parent).resolve()
else:
root = Path("~/invokeai").expanduser().resolve()
return root
class InvokeAIAppConfig(InvokeAISettings):
"""
Generate images using Stable Diffusion. Use "invokeai" to launch
@ -378,6 +199,8 @@ class InvokeAIAppConfig(InvokeAISettings):
# fmt: off
type: Literal["InvokeAI"] = "InvokeAI"
# WEB
host : str = Field(default="127.0.0.1", description="IP address to bind to", category='Web Server')
port : int = Field(default=9090, description="Port to bind to", category='Web Server')
allow_origins : List[str] = Field(default=[], description="Allowed CORS origins", category='Web Server')
@ -385,20 +208,14 @@ class InvokeAIAppConfig(InvokeAISettings):
allow_methods : List[str] = Field(default=["*"], description="Methods allowed for CORS", category='Web Server')
allow_headers : List[str] = Field(default=["*"], description="Headers allowed for CORS", category='Web Server')
# FEATURES
esrgan : bool = Field(default=True, description="Enable/disable upscaling code", category='Features')
internet_available : bool = Field(default=True, description="If true, attempt to download models on the fly; otherwise only use local models", category='Features')
log_tokenization : bool = Field(default=False, description="Enable logging of parsed prompt tokens.", category='Features')
patchmatch : bool = Field(default=True, description="Enable/disable patchmatch inpaint code", category='Features')
ignore_missing_core_models : bool = Field(default=False, description='Ignore missing models in models/core/convert', category='Features')
always_use_cpu : bool = Field(default=False, description="If true, use the CPU for rendering even if a GPU is available.", category='Memory/Performance')
free_gpu_mem : bool = Field(default=False, description="If true, purge model from GPU after each generation.", category='Memory/Performance')
max_cache_size : float = Field(default=6.0, gt=0, description="Maximum memory amount used by model cache for rapid switching", category='Memory/Performance')
max_vram_cache_size : float = Field(default=2.75, ge=0, description="Amount of VRAM reserved for model storage", category='Memory/Performance')
precision : Literal['auto', 'float16', 'float32', 'autocast'] = Field(default='auto', description='Floating point precision', category='Memory/Performance')
sequential_guidance : bool = Field(default=False, description="Whether to calculate guidance in serial instead of in parallel, lowering memory requirements", category='Memory/Performance')
xformers_enabled : bool = Field(default=True, description="Enable/disable memory-efficient attention", category='Memory/Performance')
tiled_decode : bool = Field(default=False, description="Whether to enable tiled VAE decode (reduces memory consumption with some performance penalty)", category='Memory/Performance')
# PATHS
root : Path = Field(default=None, description='InvokeAI runtime root directory', category='Paths')
autoimport_dir : Path = Field(default='autoimport', description='Path to a directory of models files to be imported on startup.', category='Paths')
lora_dir : Path = Field(default=None, description='Path to a directory of LoRA/LyCORIS models to be imported on startup.', category='Paths')
@ -409,16 +226,43 @@ class InvokeAIAppConfig(InvokeAISettings):
legacy_conf_dir : Path = Field(default='configs/stable-diffusion', description='Path to directory of legacy checkpoint config files', category='Paths')
db_dir : Path = Field(default='databases', description='Path to InvokeAI databases directory', category='Paths')
outdir : Path = Field(default='outputs', description='Default folder for output images', category='Paths')
from_file : Path = Field(default=None, description='Take command input from the indicated file (command-line client only)', category='Paths')
use_memory_db : bool = Field(default=False, description='Use in-memory database for storing image metadata', category='Paths')
ignore_missing_core_models : bool = Field(default=False, description='Ignore missing models in models/core/convert', category='Features')
from_file : Path = Field(default=None, description='Take command input from the indicated file (command-line client only)', category='Paths')
# LOGGING
log_handlers : List[str] = Field(default=["console"], description='Log handler. Valid options are "console", "file=<path>", "syslog=path|address:host:port", "http=<url>"', category="Logging")
# note - would be better to read the log_format values from logging.py, but this creates circular dependencies issues
log_format : Literal['plain', 'color', 'syslog', 'legacy'] = Field(default="color", description='Log format. Use "plain" for text-only, "color" for colorized output, "legacy" for 2.3-style logging and "syslog" for syslog-style', category="Logging")
log_level : Literal["debug", "info", "warning", "error", "critical"] = Field(default="info", description="Emit logging messages at this level or higher", category="Logging")
dev_reload : bool = Field(default=False, description="Automatically reload when Python sources are changed.", category="Development")
version : bool = Field(default=False, description="Show InvokeAI version and exit", category="Other")
# CACHE
ram : Union[float, Literal["auto"]] = Field(default=6.0, gt=0, description="Maximum memory amount used by model cache for rapid switching (floating point number or 'auto')", category="Model Cache", )
vram : Union[float, Literal["auto"]] = Field(default=0.25, ge=0, description="Amount of VRAM reserved for model storage (floating point number or 'auto')", category="Model Cache", )
lazy_offload : bool = Field(default=True, description="Keep models in VRAM until their space is needed", category="Model Cache", )
# DEVICE
device : Literal[tuple(["auto", "cpu", "cuda", "cuda:1", "mps"])] = Field(default="auto", description="Generation device", category="Device", )
precision: Literal[tuple(["auto", "float16", "float32", "autocast"])] = Field(default="auto", description="Floating point precision", category="Device", )
# GENERATION
sequential_guidance : bool = Field(default=False, description="Whether to calculate guidance in serial instead of in parallel, lowering memory requirements", category="Generation", )
attention_type : Literal[tuple(["auto", "normal", "xformers", "sliced", "torch-sdp"])] = Field(default="auto", description="Attention type", category="Generation", )
attention_slice_size: Literal[tuple(["auto", "balanced", "max", 1, 2, 3, 4, 5, 6, 7, 8])] = Field(default="auto", description='Slice size, valid when attention_type=="sliced"', category="Generation", )
force_tiled_decode: bool = Field(default=False, description="Whether to enable tiled VAE decode (reduces memory consumption with some performance penalty)", category="Generation",)
# DEPRECATED FIELDS - STILL HERE IN ORDER TO OBTAN VALUES FROM PRE-3.1 CONFIG FILES
always_use_cpu : bool = Field(default=False, description="If true, use the CPU for rendering even if a GPU is available.", category='Memory/Performance')
free_gpu_mem : Optional[bool] = Field(default=None, description="If true, purge model from GPU after each generation.", category='Memory/Performance')
max_cache_size : Optional[float] = Field(default=None, gt=0, description="Maximum memory amount used by model cache for rapid switching", category='Memory/Performance')
max_vram_cache_size : Optional[float] = Field(default=None, ge=0, description="Amount of VRAM reserved for model storage", category='Memory/Performance')
xformers_enabled : bool = Field(default=True, description="Enable/disable memory-efficient attention", category='Memory/Performance')
tiled_decode : bool = Field(default=False, description="Whether to enable tiled VAE decode (reduces memory consumption with some performance penalty)", category='Memory/Performance')
# See InvokeAIAppConfig subclass below for CACHE and DEVICE categories
# fmt: on
class Config:
@ -541,11 +385,6 @@ class InvokeAIAppConfig(InvokeAISettings):
"""Return true if precision set to float32"""
return self.precision == "float32"
@property
def disable_xformers(self) -> bool:
"""Return true if xformers_enabled is false"""
return not self.xformers_enabled
@property
def try_patchmatch(self) -> bool:
"""Return true if patchmatch true"""
@ -561,6 +400,27 @@ class InvokeAIAppConfig(InvokeAISettings):
"""invisible watermark node is always active and disabled from Web UIe"""
return True
@property
def ram_cache_size(self) -> float:
return self.max_cache_size or self.ram
@property
def vram_cache_size(self) -> float:
return self.max_vram_cache_size or self.vram
@property
def use_cpu(self) -> bool:
return self.always_use_cpu or self.device == "cpu"
@property
def disable_xformers(self) -> bool:
"""
Return true if enable_xformers is false (reversed logic)
and attention type is not set to xformers.
"""
disabled_in_config = not self.xformers_enabled
return disabled_in_config and self.attention_type != "xformers"
@staticmethod
def find_root() -> Path:
"""
@ -570,19 +430,19 @@ class InvokeAIAppConfig(InvokeAISettings):
return _find_root()
class PagingArgumentParser(argparse.ArgumentParser):
"""
A custom ArgumentParser that uses pydoc to page its output.
It also supports reading defaults from an init file.
"""
def print_help(self, file=None):
text = self.format_help()
pydoc.pager(text)
def get_invokeai_config(**kwargs) -> InvokeAIAppConfig:
"""
Legacy function which returns InvokeAIAppConfig.get_config()
"""
return InvokeAIAppConfig.get_config(**kwargs)
def _find_root() -> Path:
venv = Path(os.environ.get("VIRTUAL_ENV") or ".")
if os.environ.get("INVOKEAI_ROOT"):
root = Path(os.environ["INVOKEAI_ROOT"])
elif any([(venv.parent / x).exists() for x in [INIT_FILE, LEGACY_INIT_FILE]]):
root = (venv.parent).resolve()
else:
root = Path("~/invokeai").expanduser().resolve()
return root

18
invokeai/app/services/default_graphs.py
View File

@ -17,9 +17,9 @@ def create_text_to_image() -> LibraryGraph:
description="Converts text to an image",
graph=Graph(
nodes={
"width": IntegerInvocation(id="width", a=512),
"height": IntegerInvocation(id="height", a=512),
"seed": IntegerInvocation(id="seed", a=-1),
"width": IntegerInvocation(id="width", value=512),
"height": IntegerInvocation(id="height", value=512),
"seed": IntegerInvocation(id="seed", value=-1),
"3": NoiseInvocation(id="3"),
"4": CompelInvocation(id="4"),
"5": CompelInvocation(id="5"),
@ -29,15 +29,15 @@ def create_text_to_image() -> LibraryGraph:
},
edges=[
Edge(
source=EdgeConnection(node_id="width", field="a"),
source=EdgeConnection(node_id="width", field="value"),
destination=EdgeConnection(node_id="3", field="width"),
),
Edge(
source=EdgeConnection(node_id="height", field="a"),
source=EdgeConnection(node_id="height", field="value"),
destination=EdgeConnection(node_id="3", field="height"),
),
Edge(
source=EdgeConnection(node_id="seed", field="a"),
source=EdgeConnection(node_id="seed", field="value"),
destination=EdgeConnection(node_id="3", field="seed"),
),
Edge(
@ -65,9 +65,9 @@ def create_text_to_image() -> LibraryGraph:
exposed_inputs=[
ExposedNodeInput(node_path="4", field="prompt", alias="positive_prompt"),
ExposedNodeInput(node_path="5", field="prompt", alias="negative_prompt"),
ExposedNodeInput(node_path="width", field="a", alias="width"),
ExposedNodeInput(node_path="height", field="a", alias="height"),
ExposedNodeInput(node_path="seed", field="a", alias="seed"),
ExposedNodeInput(node_path="width", field="value", alias="width"),
ExposedNodeInput(node_path="height", field="value", alias="height"),
ExposedNodeInput(node_path="seed", field="value", alias="seed"),
],
exposed_outputs=[ExposedNodeOutput(node_path="8", field="image", alias="image")],
)

116
invokeai/app/services/invocation_stats.py
View File

@ -49,9 +49,36 @@ from invokeai.backend.model_management.model_cache import CacheStats
GIG = 1073741824
@dataclass
class NodeStats:
"""Class for tracking execution stats of an invocation node"""
calls: int = 0
time_used: float = 0.0 # seconds
max_vram: float = 0.0 # GB
cache_hits: int = 0
cache_misses: int = 0
cache_high_watermark: int = 0
@dataclass
class NodeLog:
"""Class for tracking node usage"""
# {node_type => NodeStats}
nodes: Dict[str, NodeStats] = field(default_factory=dict)
class InvocationStatsServiceBase(ABC):
"Abstract base class for recording node memory/time performance statistics"
graph_execution_manager: ItemStorageABC["GraphExecutionState"]
# {graph_id => NodeLog}
_stats: Dict[str, NodeLog]
_cache_stats: Dict[str, CacheStats]
ram_used: float
ram_changed: float
@abstractmethod
def __init__(self, graph_execution_manager: ItemStorageABC["GraphExecutionState"]):
"""
@ -94,8 +121,6 @@ class InvocationStatsServiceBase(ABC):
invocation_type: str,
time_used: float,
vram_used: float,
ram_used: float,
ram_changed: float,
):
"""
Add timing information on execution of a node. Usually
@ -104,8 +129,6 @@ class InvocationStatsServiceBase(ABC):
:param invocation_type: String literal type of the node
:param time_used: Time used by node's exection (sec)
:param vram_used: Maximum VRAM used during exection (GB)
:param ram_used: Current RAM available (GB)
:param ram_changed: Change in RAM usage over course of the run (GB)
"""
pass
@ -116,25 +139,19 @@ class InvocationStatsServiceBase(ABC):
"""
pass
@abstractmethod
def update_mem_stats(
self,
ram_used: float,
ram_changed: float,
):
"""
Update the collector with RAM memory usage info.
@dataclass
class NodeStats:
"""Class for tracking execution stats of an invocation node"""
calls: int = 0
time_used: float = 0.0 # seconds
max_vram: float = 0.0 # GB
cache_hits: int = 0
cache_misses: int = 0
cache_high_watermark: int = 0
@dataclass
class NodeLog:
"""Class for tracking node usage"""
# {node_type => NodeStats}
nodes: Dict[str, NodeStats] = field(default_factory=dict)
:param ram_used: How much RAM is currently in use.
:param ram_changed: How much RAM changed since last generation.
"""
pass
class InvocationStatsService(InvocationStatsServiceBase):
@ -152,12 +169,12 @@ class InvocationStatsService(InvocationStatsServiceBase):
class StatsContext:
"""Context manager for collecting statistics."""
invocation: BaseInvocation = None
collector: "InvocationStatsServiceBase" = None
graph_id: str = None
start_time: int = 0
ram_used: int = 0
model_manager: ModelManagerService = None
invocation: BaseInvocation
collector: "InvocationStatsServiceBase"
graph_id: str
start_time: float
ram_used: int
model_manager: ModelManagerService
def __init__(
self,
@ -170,7 +187,7 @@ class InvocationStatsService(InvocationStatsServiceBase):
self.invocation = invocation
self.collector = collector
self.graph_id = graph_id
self.start_time = 0
self.start_time = 0.0
self.ram_used = 0
self.model_manager = model_manager
@ -191,7 +208,7 @@ class InvocationStatsService(InvocationStatsServiceBase):
)
self.collector.update_invocation_stats(
graph_id=self.graph_id,
invocation_type=self.invocation.type,
invocation_type=self.invocation.type, # type: ignore - `type` is not on the `BaseInvocation` model, but *is* on all invocations
time_used=time.time() - self.start_time,
vram_used=torch.cuda.max_memory_allocated() / GIG if torch.cuda.is_available() else 0.0,
)
@ -202,11 +219,6 @@ class InvocationStatsService(InvocationStatsServiceBase):
graph_execution_state_id: str,
model_manager: ModelManagerService,
) -> StatsContext:
"""
Return a context object that will capture the statistics.
:param invocation: BaseInvocation object from the current graph.
:param graph_execution_state: GraphExecutionState object from the current session.
"""
if not self._stats.get(graph_execution_state_id): # first time we're seeing this
self._stats[graph_execution_state_id] = NodeLog()
self._cache_stats[graph_execution_state_id] = CacheStats()
@ -217,7 +229,6 @@ class InvocationStatsService(InvocationStatsServiceBase):
self._stats = {}
def reset_stats(self, graph_execution_id: str):
"""Zero the statistics for the indicated graph."""
try:
self._stats.pop(graph_execution_id)
except KeyError:
@ -228,12 +239,6 @@ class InvocationStatsService(InvocationStatsServiceBase):
ram_used: float,
ram_changed: float,
):
"""
Update the collector with RAM memory usage info.
:param ram_used: How much RAM is currently in use.
:param ram_changed: How much RAM changed since last generation.
"""
self.ram_used = ram_used
self.ram_changed = ram_changed
@ -244,16 +249,6 @@ class InvocationStatsService(InvocationStatsServiceBase):
time_used: float,
vram_used: float,
):
"""
Add timing information on execution of a node. Usually
used internally.
:param graph_id: ID of the graph that is currently executing
:param invocation_type: String literal type of the node
:param time_used: Time used by node's exection (sec)
:param vram_used: Maximum VRAM used during exection (GB)
:param ram_used: Current RAM available (GB)
:param ram_changed: Change in RAM usage over course of the run (GB)
"""
if not self._stats[graph_id].nodes.get(invocation_type):
self._stats[graph_id].nodes[invocation_type] = NodeStats()
stats = self._stats[graph_id].nodes[invocation_type]
@ -262,14 +257,15 @@ class InvocationStatsService(InvocationStatsServiceBase):
stats.max_vram = max(stats.max_vram, vram_used)
def log_stats(self):
"""
Send the statistics to the system logger at the info level.
Stats will only be printed when the execution of the graph
is complete.
"""
completed = set()
errored = set()
for graph_id, node_log in self._stats.items():
current_graph_state = self.graph_execution_manager.get(graph_id)
try:
current_graph_state = self.graph_execution_manager.get(graph_id)
except Exception:
errored.add(graph_id)
continue
if not current_graph_state.is_complete():
continue
@ -302,3 +298,7 @@ class InvocationStatsService(InvocationStatsServiceBase):
for graph_id in completed:
del self._stats[graph_id]
del self._cache_stats[graph_id]
for graph_id in errored:
del self._stats[graph_id]
del self._cache_stats[graph_id]

4
invokeai/app/services/model_manager_service.py
View File

@ -330,8 +330,8 @@ class ModelManagerService(ModelManagerServiceBase):
# configuration value. If present, then the
# cache size is set to 2.5 GB times
# the number of max_loaded_models. Otherwise
# use new `max_cache_size` config setting
max_cache_size = config.max_cache_size if hasattr(config, "max_cache_size") else config.max_loaded_models * 2.5
# use new `ram_cache_size` config setting
max_cache_size = config.ram_cache_size
logger.debug(f"Maximum RAM cache size: {max_cache_size} GiB")