Skip to content

[bugfix] reduce float value error when adding noise #9004

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

Merged
merged 5 commits into from
Oct 21, 2024
Merged

[bugfix] reduce float value error when adding noise #9004

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
11547f0
Update train_controlnet.py
gameofdimension Jul 29, 2024
eff5bcd
Update train_controlnet_sdxl.py
gameofdimension Jul 29, 2024
d669244
Merge branch 'main' into main
sayakpaul Sep 15, 2024
b814df7
Merge branch 'main' into main
gameofdimension Oct 16, 2024
d761630
style
yiyixuxu Oct 21, 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
4 changes: 3 additions & 1 deletion examples/controlnet/train_controlnet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1048,7 +1048,9 @@ def load_model_hook(models, input_dir):

# Add noise to the latents according to the noise magnitude at each timestep
# (this is the forward diffusion process)
noisy_latents = noise_scheduler.add_noise(latents, noise, timesteps)
noisy_latents = noise_scheduler.add_noise(latents.float(), noise.float(), timesteps).to(
dtype=weight_dtype
)

# Get the text embedding for conditioning
encoder_hidden_states = text_encoder(batch["input_ids"], return_dict=False)[0]
Expand Down
4 changes: 3 additions & 1 deletion examples/controlnet/train_controlnet_sdxl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1210,7 +1210,9 @@ def compute_embeddings(batch, proportion_empty_prompts, text_encoders, tokenizer

# Add noise to the latents according to the noise magnitude at each timestep
# (this is the forward diffusion process)
noisy_latents = noise_scheduler.add_noise(latents, noise, timesteps)
noisy_latents = noise_scheduler.add_noise(latents.float(), noise.float(), timesteps).to(
dtype=weight_dtype
)

# ControlNet conditioning.
controlnet_image = batch["conditioning_pixel_values"].to(dtype=weight_dtype)
Expand Down
Loading