Skip to content

Added explicit memory management during the VAE decode process. #7587

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 3 commits into
base: master
Choose a base branch
from
+15 −7
Open

Added explicit memory management during the VAE decode process. #7587

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
8e1d708
Added explicit memory management during the VAE decode process.
KBRASK Apr 13, 2025
4d4e6a0
Merge branch 'comfyanonymous:master' into master
KBRASK Apr 20, 2025
0264149
To enable support for non-CUDA devices, substitute torch.cuda.empty_c…
KBRASK Apr 20, 2025
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
22 changes: 15 additions & 7 deletions comfy/ldm/modules/diffusionmodules/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -714,13 +714,21 @@ def forward(self, z, **kwargs):
h = self.mid.block_2(h, temb, **kwargs)

# upsampling
for i_level in reversed(range(self.num_resolutions)):
for i_block in range(self.num_res_blocks+1):
h = self.up[i_level].block[i_block](h, temb, **kwargs)
if len(self.up[i_level].attn) > 0:
h = self.up[i_level].attn[i_block](h, **kwargs)
if i_level != 0:
h = self.up[i_level].upsample(h)
with torch.no_grad():
for i_level in reversed(range(self.num_resolutions)):
for i_block in range(self.num_res_blocks + 1):
h_new = self.up[i_level].block[i_block](h, temb, **kwargs)
if len(self.up[i_level].attn) > 0:
h_new = self.up[i_level].attn[i_block](h_new, **kwargs)
del h
model_management.soft_empty_cache()
h = h_new

if i_level != 0:
h_new = self.up[i_level].upsample(h)
del h
model_management.soft_empty_cache()
h = h_new

# end
if self.give_pre_end:
Expand Down