Skip to content

Fix overaggressive caching for latents with the same pixel density but different dimensions, minor fix for headless mode #251

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Fix overaggressive caching for latents with the same pixel density but different dimensions, minor fix for headless mode #251

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fe6b232
only cleanup where prompt graph exists
curiousjp Oct 6, 2024
046fcc5
added latent shape to cache key
curiousjp Oct 14, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 6 additions & 3 deletions efficiency_nodes.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,7 +134,8 @@ def efficientloader(self, ckpt_name, vae_name, clip_skip, lora_name, lora_model_
ascore=None, prompt=None, my_unique_id=None, loader_type="regular"):

# Clean globally stored objects
globals_cleanup(prompt)
if prompt:
globals_cleanup(prompt)

# Create Empty Latent
latent = torch.zeros([batch_size, 4, empty_latent_height // 8, empty_latent_width // 8]).cpu()
Expand Down Expand Up @@ -517,6 +518,7 @@ def calculate_sigmas(model_sampling, scheduler_name: str, steps):
# ------------------------------------------------------------------------------------------------------
# Store run parameters as strings. Load previous stored samples if all parameters match.
latent_image_hash = tensor_to_hash(latent_image["samples"])
latent_image_shape = latent_image["samples"].shape
positive_hash = tensor_to_hash(positive[0][0])
negative_hash = tensor_to_hash(negative[0][0])
refiner_positive_hash = tensor_to_hash(refiner_positive[0][0]) if refiner_positive is not None else None
Expand All @@ -527,7 +529,7 @@ def calculate_sigmas(model_sampling, scheduler_name: str, steps):
else [original_model_str]

parameters = [model_identifier] + [seed, steps, cfg, sampler_name, scheduler, positive_hash, negative_hash,
latent_image_hash, denoise, sampler_type, add_noise, start_at_step,
latent_image_hash, latent_image_shape, denoise, sampler_type, add_noise, start_at_step,
end_at_step, return_with_leftover_noise, refiner_model, refiner_positive_hash,
refiner_negative_hash, rng_source, cfg_denoiser, add_seed_noise, m_seed, m_weight]

Expand Down Expand Up @@ -722,7 +724,8 @@ def calculate_sigmas(model_sampling, scheduler_name: str, steps):

# ---------------------------------------------------------------------------------------------------------------
# Clean globally stored objects of non-existant nodes
globals_cleanup(prompt)
if prompt:
globals_cleanup(prompt)

# ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
# If not XY Plotting
Expand Down