mirror of
https://github.com/invoke-ai/InvokeAI
synced 2024-08-30 20:32:17 +00:00
c238a7f18b
Upgrade pydantic and fastapi to latest.
- pydantic~=2.4.2
- fastapi~=103.2
- fastapi-events~=0.9.1
**Big Changes**
There are a number of logic changes needed to support pydantic v2. Most changes are very simple, like using the new methods to serialized and deserialize models, but there are a few more complex changes.
**Invocations**
The biggest change relates to invocation creation, instantiation and validation.
Because pydantic v2 moves all validation logic into the rust pydantic-core, we may no longer directly stick our fingers into the validation pie.
Previously, we (ab)used models and fields to allow invocation fields to be optional at instantiation, but required when `invoke()` is called. We directly manipulated the fields and invocation models when calling `invoke()`.
With pydantic v2, this is much more involved. Changes to the python wrapper do not propagate down to the rust validation logic - you have to rebuild the model. This causes problem with concurrent access to the invocation classes and is not a free operation.
This logic has been totally refactored and we do not need to change the model any more. The details are in `baseinvocation.py`, in the `InputField` function and `BaseInvocation.invoke_internal()` method.
In the end, this implementation is cleaner.
**Invocation Fields**
In pydantic v2, you can no longer directly add or remove fields from a model.
Previously, we did this to add the `type` field to invocations.
**Invocation Decorators**
With pydantic v2, we instead use the imperative `create_model()` API to create a new model with the additional field. This is done in `baseinvocation.py` in the `invocation()` wrapper.
A similar technique is used for `invocation_output()`.
**Minor Changes**
There are a number of minor changes around the pydantic v2 models API.
**Protected `model_` Namespace**
All models' pydantic-provided methods and attributes are prefixed with `model_` and this is considered a protected namespace. This causes some conflict, because "model" means something to us, and we have a ton of pydantic models with attributes starting with "model_".
Forunately, there are no direct conflicts. However, in any pydantic model where we define an attribute or method that starts with "model_", we must tell set the protected namespaces to an empty tuple.
```py
class IPAdapterModelField(BaseModel):
model_name: str = Field(description="Name of the IP-Adapter model")
base_model: BaseModelType = Field(description="Base model")
model_config = ConfigDict(protected_namespaces=())
```
**Model Serialization**
Pydantic models no longer have `Model.dict()` or `Model.json()`.
Instead, we use `Model.model_dump()` or `Model.model_dump_json()`.
**Model Deserialization**
Pydantic models no longer have `Model.parse_obj()` or `Model.parse_raw()`, and there are no `parse_raw_as()` or `parse_obj_as()` functions.
Instead, you need to create a `TypeAdapter` object to parse python objects or JSON into a model.
```py
adapter_graph = TypeAdapter(Graph)
deserialized_graph_from_json = adapter_graph.validate_json(graph_json)
deserialized_graph_from_dict = adapter_graph.validate_python(graph_dict)
```
**Field Customisation**
Pydantic `Field`s no longer accept arbitrary args.
Now, you must put all additional arbitrary args in a `json_schema_extra` arg on the field.
**Schema Customisation**
FastAPI and pydantic schema generation now follows the OpenAPI version 3.1 spec.
This necessitates two changes:
- Our schema customization logic has been revised
- Schema parsing to build node templates has been revised
The specific aren't important, but this does present additional surface area for bugs.
**Performance Improvements**
Pydantic v2 is a full rewrite with a rust backend. This offers a substantial performance improvement (pydantic claims 5x to 50x depending on the task). We'll notice this the most during serialization and deserialization of sessions/graphs, which happens very very often - a couple times per node.
I haven't done any benchmarks, but anecdotally, graph execution is much faster. Also, very larges graphs - like with massive iterators - are much, much faster.
130 lines
3.3 KiB
Python
130 lines
3.3 KiB
Python
# Copyright (c) 2023 Kyle Schouviller (https://github.com/kyle0654) & the InvokeAI Team
|
|
|
|
|
|
import torch
|
|
from pydantic import field_validator
|
|
|
|
from invokeai.app.invocations.latent import LatentsField
|
|
from invokeai.app.util.misc import SEED_MAX, get_random_seed
|
|
|
|
from ...backend.util.devices import choose_torch_device, torch_dtype
|
|
from .baseinvocation import (
|
|
BaseInvocation,
|
|
BaseInvocationOutput,
|
|
FieldDescriptions,
|
|
InputField,
|
|
InvocationContext,
|
|
OutputField,
|
|
invocation,
|
|
invocation_output,
|
|
)
|
|
|
|
"""
|
|
Utilities
|
|
"""
|
|
|
|
|
|
def get_noise(
|
|
width: int,
|
|
height: int,
|
|
device: torch.device,
|
|
seed: int = 0,
|
|
latent_channels: int = 4,
|
|
downsampling_factor: int = 8,
|
|
use_cpu: bool = True,
|
|
perlin: float = 0.0,
|
|
):
|
|
"""Generate noise for a given image size."""
|
|
noise_device_type = "cpu" if use_cpu else device.type
|
|
|
|
# limit noise to only the diffusion image channels, not the mask channels
|
|
input_channels = min(latent_channels, 4)
|
|
generator = torch.Generator(device=noise_device_type).manual_seed(seed)
|
|
|
|
noise_tensor = torch.randn(
|
|
[
|
|
1,
|
|
input_channels,
|
|
height // downsampling_factor,
|
|
width // downsampling_factor,
|
|
],
|
|
dtype=torch_dtype(device),
|
|
device=noise_device_type,
|
|
generator=generator,
|
|
).to("cpu")
|
|
|
|
return noise_tensor
|
|
|
|
|
|
"""
|
|
Nodes
|
|
"""
|
|
|
|
|
|
@invocation_output("noise_output")
|
|
class NoiseOutput(BaseInvocationOutput):
|
|
"""Invocation noise output"""
|
|
|
|
noise: LatentsField = OutputField(description=FieldDescriptions.noise)
|
|
width: int = OutputField(description=FieldDescriptions.width)
|
|
height: int = OutputField(description=FieldDescriptions.height)
|
|
|
|
|
|
def build_noise_output(latents_name: str, latents: torch.Tensor, seed: int):
|
|
return NoiseOutput(
|
|
noise=LatentsField(latents_name=latents_name, seed=seed),
|
|
width=latents.size()[3] * 8,
|
|
height=latents.size()[2] * 8,
|
|
)
|
|
|
|
|
|
@invocation(
|
|
"noise",
|
|
title="Noise",
|
|
tags=["latents", "noise"],
|
|
category="latents",
|
|
version="1.0.0",
|
|
)
|
|
class NoiseInvocation(BaseInvocation):
|
|
"""Generates latent noise."""
|
|
|
|
seed: int = InputField(
|
|
ge=0,
|
|
le=SEED_MAX,
|
|
description=FieldDescriptions.seed,
|
|
default_factory=get_random_seed,
|
|
)
|
|
width: int = InputField(
|
|
default=512,
|
|
multiple_of=8,
|
|
gt=0,
|
|
description=FieldDescriptions.width,
|
|
)
|
|
height: int = InputField(
|
|
default=512,
|
|
multiple_of=8,
|
|
gt=0,
|
|
description=FieldDescriptions.height,
|
|
)
|
|
use_cpu: bool = InputField(
|
|
default=True,
|
|
description="Use CPU for noise generation (for reproducible results across platforms)",
|
|
)
|
|
|
|
@field_validator("seed", mode="before")
|
|
def modulo_seed(cls, v):
|
|
"""Returns the seed modulo (SEED_MAX + 1) to ensure it is within the valid range."""
|
|
return v % (SEED_MAX + 1)
|
|
|
|
def invoke(self, context: InvocationContext) -> NoiseOutput:
|
|
noise = get_noise(
|
|
width=self.width,
|
|
height=self.height,
|
|
device=choose_torch_device(),
|
|
seed=self.seed,
|
|
use_cpu=self.use_cpu,
|
|
)
|
|
name = f"{context.graph_execution_state_id}__{self.id}"
|
|
context.services.latents.save(name, noise)
|
|
return build_noise_output(latents_name=name, latents=noise, seed=self.seed)
|