Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
diffusers
Commits
7e11392d
Commit
7e11392d
authored
Jul 19, 2022
by
Patrick von Platen
Browse files
fix ddpm scheduler
parent
1f49a343
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
12 additions
and
9 deletions
+12
-9
src/diffusers/pipelines/ddpm/pipeline_ddpm.py
src/diffusers/pipelines/ddpm/pipeline_ddpm.py
+2
-8
src/diffusers/schedulers/scheduling_ddpm.py
src/diffusers/schedulers/scheduling_ddpm.py
+10
-1
No files found.
src/diffusers/pipelines/ddpm/pipeline_ddpm.py
View file @
7e11392d
...
@@ -51,13 +51,7 @@ class DDPMPipeline(DiffusionPipeline):
...
@@ -51,13 +51,7 @@ class DDPMPipeline(DiffusionPipeline):
# 2. predict previous mean of image x_t-1
# 2. predict previous mean of image x_t-1
pred_prev_image
=
self
.
scheduler
.
step
(
model_output
,
t
,
image
)[
"prev_sample"
]
pred_prev_image
=
self
.
scheduler
.
step
(
model_output
,
t
,
image
)[
"prev_sample"
]
# 3. optionally sample variance
# 3. set current image to prev_image: x_t -> x_t-1
variance
=
0
image
=
pred_prev_image
if
t
>
0
:
noise
=
torch
.
randn
(
image
.
shape
,
generator
=
generator
).
to
(
image
.
device
)
variance
=
self
.
scheduler
.
get_variance
(
t
).
sqrt
()
*
noise
# 4. set current image to prev_image: x_t -> x_t-1
image
=
pred_prev_image
+
variance
return
{
"sample"
:
image
}
return
{
"sample"
:
image
}
src/diffusers/schedulers/scheduling_ddpm.py
View file @
7e11392d
...
@@ -101,7 +101,7 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
...
@@ -101,7 +101,7 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
)[::
-
1
].
copy
()
)[::
-
1
].
copy
()
self
.
set_format
(
tensor_format
=
self
.
tensor_format
)
self
.
set_format
(
tensor_format
=
self
.
tensor_format
)
def
get_variance
(
self
,
t
,
variance_type
=
None
):
def
_
get_variance
(
self
,
t
,
variance_type
=
None
):
alpha_prod_t
=
self
.
alphas_cumprod
[
t
]
alpha_prod_t
=
self
.
alphas_cumprod
[
t
]
alpha_prod_t_prev
=
self
.
alphas_cumprod
[
t
-
1
]
if
t
>
0
else
self
.
one
alpha_prod_t_prev
=
self
.
alphas_cumprod
[
t
-
1
]
if
t
>
0
else
self
.
one
...
@@ -133,6 +133,7 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
...
@@ -133,6 +133,7 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
timestep
:
int
,
timestep
:
int
,
sample
:
Union
[
torch
.
FloatTensor
,
np
.
ndarray
],
sample
:
Union
[
torch
.
FloatTensor
,
np
.
ndarray
],
predict_epsilon
=
True
,
predict_epsilon
=
True
,
generator
=
None
,
):
):
t
=
timestep
t
=
timestep
# 1. compute alphas, betas
# 1. compute alphas, betas
...
@@ -161,6 +162,14 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
...
@@ -161,6 +162,14 @@ class DDPMScheduler(SchedulerMixin, ConfigMixin):
# See formula (7) from https://arxiv.org/pdf/2006.11239.pdf
# See formula (7) from https://arxiv.org/pdf/2006.11239.pdf
pred_prev_sample
=
pred_original_sample_coeff
*
pred_original_sample
+
current_sample_coeff
*
sample
pred_prev_sample
=
pred_original_sample_coeff
*
pred_original_sample
+
current_sample_coeff
*
sample
# 6. Add noise
variance
=
0
if
t
>
0
:
noise
=
torch
.
randn
(
model_output
.
shape
,
generator
=
generator
).
to
(
model_output
.
device
)
variance
=
self
.
_get_variance
(
t
).
sqrt
()
*
noise
pred_prev_sample
=
pred_prev_sample
+
variance
return
{
"prev_sample"
:
pred_prev_sample
}
return
{
"prev_sample"
:
pred_prev_sample
}
def
add_noise
(
self
,
original_samples
,
noise
,
timesteps
):
def
add_noise
(
self
,
original_samples
,
noise
,
timesteps
):
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment