github/InvokeAI
1
1
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI.git synced 2026-02-11 19:17:20 -05:00
Code Issues Packages Projects Releases Wiki Activity

Rename ModelPatcher -> LayerPatcher to avoid conflicts with another ModelPatcher definition.

Browse Source
This commit is contained in:
Ryan Dick
2024-12-14 16:11:23 +00:00
parent 7fad4c9491
commit dd09509dbd
9 changed files with 33 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
invokeai/app/invocations/compel.py
View File

@@ -21,7 +21,7 @@ from invokeai.app.services.shared.invocation_context import InvocationContext
from invokeai.app.util.ti_utils import generate_ti_list
from invokeai.backend.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patch_raw import ModelPatchRaw
from invokeai.backend.patches.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patcher import LayerPatcher
from invokeai.backend.stable_diffusion.diffusion.conditioning_data import (
BasicConditioningInfo,
ConditioningFieldData,
@@ -82,7 +82,7 @@ class CompelInvocation(BaseInvocation):
# apply all patches while the model is on the target device
text_encoder_info.model_on_device() as (cached_weights, text_encoder),
tokenizer_info as tokenizer,
ModelPatcher.apply_model_patches(
LayerPatcher.apply_model_patches(
model=text_encoder,
patches=_lora_loader(),
prefix="lora_te_",
@@ -179,7 +179,7 @@ class SDXLPromptInvocationBase:
# apply all patches while the model is on the target device
text_encoder_info.model_on_device() as (cached_weights, text_encoder),
tokenizer_info as tokenizer,
ModelPatcher.apply_model_patches(
LayerPatcher.apply_model_patches(
text_encoder,
patches=_lora_loader(),
prefix=lora_prefix,

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

@@ -40,7 +40,7 @@ from invokeai.backend.ip_adapter.ip_adapter import IPAdapter
from invokeai.backend.model_manager import BaseModelType, ModelVariantType
from invokeai.backend.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patch_raw import ModelPatchRaw
from invokeai.backend.patches.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patcher import LayerPatcher
from invokeai.backend.stable_diffusion import PipelineIntermediateState
from invokeai.backend.stable_diffusion.denoise_context import DenoiseContext, DenoiseInputs
from invokeai.backend.stable_diffusion.diffusers_pipeline import (
@@ -1003,7 +1003,7 @@ class DenoiseLatentsInvocation(BaseInvocation):
ModelPatcher.apply_freeu(unet, self.unet.freeu_config),
SeamlessExt.static_patch_model(unet, self.unet.seamless_axes), # FIXME
# Apply the LoRA after unet has been moved to its target device for faster patching.
ModelPatcher.apply_model_patches(
LayerPatcher.apply_model_patches(
model=unet,
patches=_lora_loader(),
prefix="lora_unet_",

6
invokeai/app/invocations/flux_denoise.py
View File

@@ -50,7 +50,7 @@ from invokeai.backend.flux.text_conditioning import FluxTextConditioning
from invokeai.backend.model_manager.config import ModelFormat
from invokeai.backend.patches.lora_conversions.flux_lora_constants import FLUX_LORA_TRANSFORMER_PREFIX
from invokeai.backend.patches.model_patch_raw import ModelPatchRaw
from invokeai.backend.patches.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patcher import LayerPatcher
from invokeai.backend.stable_diffusion.diffusers_pipeline import PipelineIntermediateState
from invokeai.backend.stable_diffusion.diffusion.conditioning_data import FLUXConditioningInfo
from invokeai.backend.util.devices import TorchDevice
@@ -306,7 +306,7 @@ class FluxDenoiseInvocation(BaseInvocation, WithMetadata, WithBoard):
if config.format in [ModelFormat.Checkpoint]:
# The model is non-quantized, so we can apply the LoRA weights directly into the model.
exit_stack.enter_context(
ModelPatcher.apply_model_patches(
LayerPatcher.apply_model_patches(
model=transformer,
patches=self._lora_iterator(context),
prefix=FLUX_LORA_TRANSFORMER_PREFIX,
@@ -321,7 +321,7 @@ class FluxDenoiseInvocation(BaseInvocation, WithMetadata, WithBoard):
# The model is quantized, so apply the LoRA weights as sidecar layers. This results in slower inference,
# than directly patching the weights, but is agnostic to the quantization format.
exit_stack.enter_context(
ModelPatcher.apply_model_sidecar_patches(
LayerPatcher.apply_model_sidecar_patches(
model=transformer,
patches=self._lora_iterator(context),
prefix=FLUX_LORA_TRANSFORMER_PREFIX,

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

@@ -20,7 +20,7 @@ from invokeai.backend.flux.modules.conditioner import HFEncoder
from invokeai.backend.model_manager.config import ModelFormat
from invokeai.backend.patches.lora_conversions.flux_lora_constants import FLUX_LORA_CLIP_PREFIX
from invokeai.backend.patches.model_patch_raw import ModelPatchRaw
from invokeai.backend.patches.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patcher import LayerPatcher
from invokeai.backend.stable_diffusion.diffusion.conditioning_data import ConditioningFieldData, FLUXConditioningInfo
@@ -111,7 +111,7 @@ class FluxTextEncoderInvocation(BaseInvocation):
if clip_text_encoder_config.format in [ModelFormat.Diffusers]:
# The model is non-quantized, so we can apply the LoRA weights directly into the model.
exit_stack.enter_context(
ModelPatcher.apply_model_patches(
LayerPatcher.apply_model_patches(
model=clip_text_encoder,
patches=self._clip_lora_iterator(context),
prefix=FLUX_LORA_CLIP_PREFIX,

4
invokeai/app/invocations/sd3_text_encoder.py
View File

@@ -19,7 +19,7 @@ from invokeai.app.services.shared.invocation_context import InvocationContext
from invokeai.backend.model_manager.config import ModelFormat
from invokeai.backend.patches.lora_conversions.flux_lora_constants import FLUX_LORA_CLIP_PREFIX
from invokeai.backend.patches.model_patch_raw import ModelPatchRaw
from invokeai.backend.patches.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patcher import LayerPatcher
from invokeai.backend.stable_diffusion.diffusion.conditioning_data import ConditioningFieldData, SD3ConditioningInfo
# The SD3 T5 Max Sequence Length set based on the default in diffusers.
@@ -150,7 +150,7 @@ class Sd3TextEncoderInvocation(BaseInvocation):
if clip_text_encoder_config.format in [ModelFormat.Diffusers]:
# The model is non-quantized, so we can apply the LoRA weights directly into the model.
exit_stack.enter_context(
ModelPatcher.apply_model_patches(
LayerPatcher.apply_model_patches(
model=clip_text_encoder,
patches=self._clip_lora_iterator(context, clip_model),
prefix=FLUX_LORA_CLIP_PREFIX,

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

@@ -23,7 +23,7 @@ from invokeai.app.invocations.model import UNetField
from invokeai.app.invocations.primitives import LatentsOutput
from invokeai.app.services.shared.invocation_context import InvocationContext
from invokeai.backend.patches.model_patch_raw import ModelPatchRaw
from invokeai.backend.patches.model_patcher import ModelPatcher
from invokeai.backend.patches.model_patcher import LayerPatcher
from invokeai.backend.stable_diffusion.diffusers_pipeline import ControlNetData, PipelineIntermediateState
from invokeai.backend.stable_diffusion.multi_diffusion_pipeline import (
MultiDiffusionPipeline,
@@ -207,7 +207,7 @@ class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
with (
ExitStack() as exit_stack,
unet_info as unet,
ModelPatcher.apply_model_patches(model=unet, patches=_lora_loader(), prefix="lora_unet_"),
LayerPatcher.apply_model_patches(model=unet, patches=_lora_loader(), prefix="lora_unet_"),
):
assert isinstance(unet, UNet2DConditionModel)
latents = latents.to(device=unet.device, dtype=unet.dtype)