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

Cleaning up mistakes after rebase.

Browse Source
This commit is contained in:
user1 2023-05-11 03:27:21 -07:00 committed by Kent Keirsey
parent 7e70391c2b
commit 70ba36eefc
1 changed files with 0 additions and 66 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

66
invokeai/app/invocations/latent.py
View File

@ -329,72 +329,6 @@ class TextToLatentsInvocation(BaseInvocation):
multi_control = MultiControlNetModel(control_models)
model.control_model = multi_control
print("type of control input: ", type(self.control))
if (self.control is None):
control_model = None
control_image_field = None
control_weight = None
else:
control_model_name = self.control.control_model
control_image_field = self.control.image
control_weight = self.control.control_weight
# # loading controlnet model
# if (self.control_model is None or self.control_model==''):
# control_model = None
# else:
# FIXME: change this to dropdown menu?
# FIXME: generalize so don't have to hardcode torch_dtype and device
control_model = ControlNetModel.from_pretrained(control_model_name,
torch_dtype=torch.float16).to("cuda")
model.control_model = control_model
# loading controlnet image (currently requires pre-processed image)
control_image = (
None if control_image_field is None
else context.services.images.get(
control_image_field.image_type, control_image_field.image_name
)
)
latents_shape = noise.shape
control_height_resize = latents_shape[2] * 8
control_width_resize = latents_shape[3] * 8
# copied from old backend/txt2img.py
# FIXME: still need to test with different widths, heights, devices, dtypes
# and add in batch_size, num_images_per_prompt?
if control_image is not None:
if isinstance(control_model, ControlNetModel):
control_image = model.prepare_control_image(
image=control_image,
# do_classifier_free_guidance=do_classifier_free_guidance,
do_classifier_free_guidance=True,
width=control_width_resize,
height=control_height_resize,
# batch_size=batch_size * num_images_per_prompt,
# num_images_per_prompt=num_images_per_prompt,
device=control_model.device,
dtype=control_model.dtype,
)
elif isinstance(control_model, MultiControlNetModel):
images = []
for image_ in control_image:
image_ = model.prepare_control_image(
image=image_,
# do_classifier_free_guidance=do_classifier_free_guidance,
do_classifier_free_guidance=True,
width=control_width_resize,
height=control_height_resize,
# batch_size=batch_size * num_images_per_prompt,
# num_images_per_prompt=num_images_per_prompt,
device=control_model.device,
dtype=control_model.dtype,
)
images.append(image_)
control_image = images
# TODO: Verify the noise is the right size
result_latents, result_attention_map_saver = model.latents_from_embeddings(
latents=torch.zeros_like(noise, dtype=torch_dtype(model.device)),