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

Rename as suggested in other PRs

Browse Source 
Co-Authored-By: Ryan Dick <14897797+RyanJDick@users.noreply.github.com>
This commit is contained in:
Sergey Borisov 2024-07-23 01:01:18 +03:00
parent 42356ec866
commit 3cb13d6288
1 changed files with 22 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

44
invokeai/backend/stable_diffusion/extensions/controlnet.py
View File

@ -31,25 +31,25 @@ class ControlNetExt(ExtensionBase):
resize_mode: str, resize_mode: str,
): ):
super().__init__() super().__init__()
self.model = model self._model = model
self.image = image self._image = image
self.weight = weight self._weight = weight
self.begin_step_percent = begin_step_percent self._begin_step_percent = begin_step_percent
self.end_step_percent = end_step_percent self._end_step_percent = end_step_percent
self.control_mode = control_mode self._control_mode = control_mode
self.resize_mode = resize_mode self._resize_mode = resize_mode
self.image_tensor: Optional[torch.Tensor] = None self._image_tensor: Optional[torch.Tensor] = None
@contextmanager @contextmanager
def patch_extension(self, ctx: DenoiseContext): def patch_extension(self, ctx: DenoiseContext):
try: try:
original_processors = self.model.attn_processors original_processors = self._model.attn_processors
self.model.set_attn_processor(ctx.inputs.attention_processor_cls()) self._model.set_attn_processor(ctx.inputs.attention_processor_cls())
yield None yield None
finally: finally:
self.model.set_attn_processor(original_processors) self._model.set_attn_processor(original_processors)
@callback(ExtensionCallbackType.PRE_DENOISE_LOOP) @callback(ExtensionCallbackType.PRE_DENOISE_LOOP)
def resize_image(self, ctx: DenoiseContext): def resize_image(self, ctx: DenoiseContext):
@ -57,8 +57,8 @@ class ControlNetExt(ExtensionBase):
image_height = latent_height * LATENT_SCALE_FACTOR image_height = latent_height * LATENT_SCALE_FACTOR
image_width = latent_width * LATENT_SCALE_FACTOR image_width = latent_width * LATENT_SCALE_FACTOR
self.image_tensor = prepare_control_image( self._image_tensor = prepare_control_image(
image=self.image, image=self._image,
do_classifier_free_guidance=False, do_classifier_free_guidance=False,
width=image_width, width=image_width,
height=image_height, height=image_height,
@ -66,22 +66,22 @@ class ControlNetExt(ExtensionBase):
# num_images_per_prompt=num_images_per_prompt, # num_images_per_prompt=num_images_per_prompt,
device=ctx.latents.device, device=ctx.latents.device,
dtype=ctx.latents.dtype, dtype=ctx.latents.dtype,
control_mode=self.control_mode, control_mode=self._control_mode,
resize_mode=self.resize_mode, resize_mode=self._resize_mode,
) )
@callback(ExtensionCallbackType.PRE_UNET) @callback(ExtensionCallbackType.PRE_UNET)
def pre_unet_step(self, ctx: DenoiseContext): def pre_unet_step(self, ctx: DenoiseContext):
# skip if model not active in current step # skip if model not active in current step
total_steps = len(ctx.inputs.timesteps) total_steps = len(ctx.inputs.timesteps)
first_step = math.floor(self.begin_step_percent * total_steps) first_step = math.floor(self._begin_step_percent * total_steps)
last_step = math.ceil(self.end_step_percent * total_steps) last_step = math.ceil(self._end_step_percent * total_steps)
if ctx.step_index < first_step or ctx.step_index > last_step: if ctx.step_index < first_step or ctx.step_index > last_step:
return return
# convert mode to internal flags # convert mode to internal flags
soft_injection = self.control_mode in ["more_prompt", "more_control"] soft_injection = self._control_mode in ["more_prompt", "more_control"]
cfg_injection = self.control_mode in ["more_control", "unbalanced"] cfg_injection = self._control_mode in ["more_control", "unbalanced"]
# no negative conditioning in cfg_injection mode # no negative conditioning in cfg_injection mode
if cfg_injection: if cfg_injection:
@ -117,7 +117,7 @@ class ControlNetExt(ExtensionBase):
total_steps = len(ctx.inputs.timesteps) total_steps = len(ctx.inputs.timesteps)
model_input = ctx.latent_model_input model_input = ctx.latent_model_input
image_tensor = self.image_tensor image_tensor = self._image_tensor
if conditioning_mode == ConditioningMode.Both: if conditioning_mode == ConditioningMode.Both:
model_input = torch.cat([model_input] * 2) model_input = torch.cat([model_input] * 2)
image_tensor = torch.cat([image_tensor] * 2) image_tensor = torch.cat([image_tensor] * 2)
@ -134,7 +134,7 @@ class ControlNetExt(ExtensionBase):
ctx.inputs.conditioning_data.to_unet_kwargs(cn_unet_kwargs, conditioning_mode=conditioning_mode) ctx.inputs.conditioning_data.to_unet_kwargs(cn_unet_kwargs, conditioning_mode=conditioning_mode)
# get static weight, or weight corresponding to current step # get static weight, or weight corresponding to current step
weight = self.weight weight = self._weight
if isinstance(weight, list): if isinstance(weight, list):
weight = weight[ctx.step_index] weight = weight[ctx.step_index]
@ -144,7 +144,7 @@ class ControlNetExt(ExtensionBase):
tmp_kwargs.pop("down_intrablock_additional_residuals", None) tmp_kwargs.pop("down_intrablock_additional_residuals", None)
# controlnet(s) inference # controlnet(s) inference
down_samples, mid_sample = self.model( down_samples, mid_sample = self._model(
controlnet_cond=image_tensor, controlnet_cond=image_tensor,
conditioning_scale=weight, # controlnet specific, NOT the guidance scale conditioning_scale=weight, # controlnet specific, NOT the guidance scale
guess_mode=soft_injection, # this is still called guess_mode in diffusers ControlNetModel guess_mode=soft_injection, # this is still called guess_mode in diffusers ControlNetModel