Skip to content
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

Jump to bottom

Option to normalize latent interpolation images #438

Merged
merged 3 commits into from
Dec 8, 2020
Merged

Option to normalize latent interpolation images #438

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
a45a64b
add option to normalize latent interpolation images
teddykoker Dec 8, 2020
4074654
linspace
teddykoker Dec 8, 2020
2e3199b
update
ananyahjha93 Dec 8, 2020
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
19 changes: 15 additions & 4 deletions pl_bolts/callbacks/variational.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,5 @@
import math
import numpy as np

import torch
from pytorch_lightning.callbacks import Callback
Expand Down Expand Up @@ -26,20 +27,30 @@ class LatentDimInterpolator(Callback):
"""

def __init__(
self, interpolate_epoch_interval: int = 20, range_start: int = -5, range_end: int = 5, num_samples: int = 2
self,
interpolate_epoch_interval: int = 20,
range_start: int = -5,
range_end: int = 5,
steps: int = 11,
num_samples: int = 2,
normalize: bool = True,
):
"""
Args:
interpolate_epoch_interval: default 20
range_start: default -5
range_end: default 5
steps: number of step between start and end
num_samples: default 2
normalize: default True (change image to (0, 1) range)
"""
super().__init__()
self.interpolate_epoch_interval = interpolate_epoch_interval
self.range_start = range_start
self.range_end = range_end
self.num_samples = num_samples
self.normalize = normalize
self.steps = steps

def on_epoch_end(self, trainer, pl_module):
if (trainer.current_epoch + 1) % self.interpolate_epoch_interval == 0:
Expand All @@ -48,16 +59,16 @@ def on_epoch_end(self, trainer, pl_module):

num_images = (self.range_end - self.range_start) ** 2
num_rows = int(math.sqrt(num_images))
grid = torchvision.utils.make_grid(images, nrow=num_rows)
grid = torchvision.utils.make_grid(images, nrow=num_rows, normalize=self.normalize)
str_title = f'{pl_module.__class__.__name__}_latent_space'
trainer.logger.experiment.add_image(str_title, grid, global_step=trainer.global_step)

def interpolate_latent_space(self, pl_module, latent_dim):
images = []
with torch.no_grad():
pl_module.eval()
for z1 in range(self.range_start, self.range_end, 1):
for z2 in range(self.range_start, self.range_end, 1):
for z1 in np.linspace(self.range_start, self.range_end, self.steps):
for z2 in np.linspace(self.range_start, self.range_end, self.steps):
# set all dims to zero
z = torch.zeros(self.num_samples, latent_dim, device=pl_module.device)

Expand Down