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

(minor) Address a few small TODOs.

Browse Source
This commit is contained in:
Ryan Dick 2024-06-19 11:07:35 -04:00 committed by Kent Keirsey
parent bd3b6ca11b
commit 8a43656cf9
1 changed files with 8 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
invokeai/app/invocations/tiled_multi_diffusion_denoise_latents.py
View File

@ -55,15 +55,14 @@ def crop_controlnet_data(control_data: ControlNetData, latent_region: TBLR) -> C
title="Tiled Multi-Diffusion Denoise Latents", title="Tiled Multi-Diffusion Denoise Latents",
tags=["upscale", "denoise"], tags=["upscale", "denoise"],
category="latents", category="latents",
# TODO(ryand): Reset to 1.0.0 right before release.
version="1.0.0", version="1.0.0",
) )
class TiledMultiDiffusionDenoiseLatents(BaseInvocation): class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
"""Tiled Multi-Diffusion denoising. """Tiled Multi-Diffusion denoising.
This node handles automatically tiling the input image. Future iterations of This node handles automatically tiling the input image, and is primarily intended for global refinement of images
this node should allow the user to specify custom regions with different parameters for each region to harness the in tiled upscaling workflows. Future Multi-Diffusion nodes should allow the user to specify custom regions with
full power of Multi-Diffusion. different parameters for each region to harness the full power of Multi-Diffusion.
This node has a similar interface to the `DenoiseLatents` node, but it has a reduced feature set (no IP-Adapter, This node has a similar interface to the `DenoiseLatents` node, but it has a reduced feature set (no IP-Adapter,
T2I-Adapter, masking, etc.). T2I-Adapter, masking, etc.).
@ -85,10 +84,8 @@ class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
description=FieldDescriptions.latents, description=FieldDescriptions.latents,
input=Input.Connection, input=Input.Connection,
) )
# TODO(ryand): Add multiple-of validation. tile_height: int = InputField(default=64, gt=0, description="Height of the tiles in latent space.")
# TODO(ryand): Smaller defaults might make more sense. tile_width: int = InputField(default=64, gt=0, description="Width of the tiles in latent space.")
tile_height: int = InputField(default=112, gt=0, description="Height of the tiles in latent space.")
tile_width: int = InputField(default=112, gt=0, description="Width of the tiles in latent space.")
tile_min_overlap: int = InputField( tile_min_overlap: int = InputField(
default=16, default=16,
gt=0, gt=0,
@ -97,9 +94,8 @@ class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
) )
steps: int = InputField(default=18, gt=0, description=FieldDescriptions.steps) steps: int = InputField(default=18, gt=0, description=FieldDescriptions.steps)
cfg_scale: float | list[float] = InputField(default=6.0, description=FieldDescriptions.cfg_scale, title="CFG Scale") cfg_scale: float | list[float] = InputField(default=6.0, description=FieldDescriptions.cfg_scale, title="CFG Scale")
# TODO(ryand): The default here should probably be 0.0.
denoising_start: float = InputField( denoising_start: float = InputField(
default=0.65, default=0.0,
ge=0, ge=0,
le=1, le=1,
description=FieldDescriptions.denoising_start, description=FieldDescriptions.denoising_start,
@ -150,7 +146,7 @@ class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
self.config = FakeVae.FakeVaeConfig() self.config = FakeVae.FakeVaeConfig()
return MultiDiffusionPipeline( return MultiDiffusionPipeline(
vae=FakeVae(), # TODO: oh... vae=FakeVae(),
text_encoder=None, text_encoder=None,
tokenizer=None, tokenizer=None,
unet=unet, unet=unet,
@ -166,7 +162,6 @@ class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
_, _, latent_height, latent_width = latents.shape _, _, latent_height, latent_width = latents.shape
# Calculate the tile locations to cover the latent-space image. # Calculate the tile locations to cover the latent-space image.
# TODO(ryand): Add constraints on the tile params. Is there a multiple-of constraint?
tiles = calc_tiles_min_overlap( tiles = calc_tiles_min_overlap(
image_height=latent_height, image_height=latent_height,
image_width=latent_width, image_width=latent_width,
@ -260,8 +255,8 @@ class TiledMultiDiffusionDenoiseLatents(BaseInvocation):
callback=lambda x: None, callback=lambda x: None,
) )
# TODO(ryand): I copied this from DenoiseLatentsInvocation. I'm not sure if it's actually important.
result_latents = result_latents.to("cpu") result_latents = result_latents.to("cpu")
# TODO(ryand): I copied this from DenoiseLatentsInvocation. I'm not sure if it's actually important.
TorchDevice.empty_cache() TorchDevice.empty_cache()
name = context.tensors.save(tensor=result_latents) name = context.tensors.save(tensor=result_latents)