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.
103 lines
3.5 KiB
Python
103 lines
3.5 KiB
Python
from __future__ import annotations
|
|
|
|
from contextlib import contextmanager
|
|
from typing import List, Union
|
|
|
|
import torch.nn as nn
|
|
from diffusers.models import AutoencoderKL, UNet2DConditionModel
|
|
|
|
|
|
def _conv_forward_asymmetric(self, input, weight, bias):
|
|
"""
|
|
Patch for Conv2d._conv_forward that supports asymmetric padding
|
|
"""
|
|
working = nn.functional.pad(input, self.asymmetric_padding["x"], mode=self.asymmetric_padding_mode["x"])
|
|
working = nn.functional.pad(working, self.asymmetric_padding["y"], mode=self.asymmetric_padding_mode["y"])
|
|
return nn.functional.conv2d(
|
|
working,
|
|
weight,
|
|
bias,
|
|
self.stride,
|
|
nn.modules.utils._pair(0),
|
|
self.dilation,
|
|
self.groups,
|
|
)
|
|
|
|
|
|
@contextmanager
|
|
def set_seamless(model: Union[UNet2DConditionModel, AutoencoderKL], seamless_axes: List[str]):
|
|
try:
|
|
to_restore = []
|
|
|
|
for m_name, m in model.named_modules():
|
|
if isinstance(model, UNet2DConditionModel):
|
|
if ".attentions." in m_name:
|
|
continue
|
|
|
|
if ".resnets." in m_name:
|
|
if ".conv2" in m_name:
|
|
continue
|
|
if ".conv_shortcut" in m_name:
|
|
continue
|
|
|
|
"""
|
|
if isinstance(model, UNet2DConditionModel):
|
|
if False and ".upsamplers." in m_name:
|
|
continue
|
|
|
|
if False and ".downsamplers." in m_name:
|
|
continue
|
|
|
|
if True and ".resnets." in m_name:
|
|
if True and ".conv1" in m_name:
|
|
if False and "down_blocks" in m_name:
|
|
continue
|
|
if False and "mid_block" in m_name:
|
|
continue
|
|
if False and "up_blocks" in m_name:
|
|
continue
|
|
|
|
if True and ".conv2" in m_name:
|
|
continue
|
|
|
|
if True and ".conv_shortcut" in m_name:
|
|
continue
|
|
|
|
if True and ".attentions." in m_name:
|
|
continue
|
|
|
|
if False and m_name in ["conv_in", "conv_out"]:
|
|
continue
|
|
"""
|
|
|
|
if isinstance(m, (nn.Conv2d, nn.ConvTranspose2d)):
|
|
m.asymmetric_padding_mode = {}
|
|
m.asymmetric_padding = {}
|
|
m.asymmetric_padding_mode["x"] = "circular" if ("x" in seamless_axes) else "constant"
|
|
m.asymmetric_padding["x"] = (
|
|
m._reversed_padding_repeated_twice[0],
|
|
m._reversed_padding_repeated_twice[1],
|
|
0,
|
|
0,
|
|
)
|
|
m.asymmetric_padding_mode["y"] = "circular" if ("y" in seamless_axes) else "constant"
|
|
m.asymmetric_padding["y"] = (
|
|
0,
|
|
0,
|
|
m._reversed_padding_repeated_twice[2],
|
|
m._reversed_padding_repeated_twice[3],
|
|
)
|
|
|
|
to_restore.append((m, m._conv_forward))
|
|
m._conv_forward = _conv_forward_asymmetric.__get__(m, nn.Conv2d)
|
|
|
|
yield
|
|
|
|
finally:
|
|
for module, orig_conv_forward in to_restore:
|
|
module._conv_forward = orig_conv_forward
|
|
if hasattr(module, "asymmetric_padding_mode"):
|
|
del module.asymmetric_padding_mode
|
|
if hasattr(module, "asymmetric_padding"):
|
|
del module.asymmetric_padding
|