DL-Art-School/codes/scripts/srflow_latent_space_playground.py

268 lines
11 KiB
Python
Raw Normal View History

import argparse
import logging
import math
import os
from glob import glob
import torch
import torch.nn.functional as F
import torchvision
from PIL import Image
from tqdm import tqdm
import utils.options as option
import utils
2022-03-16 18:05:56 +00:00
from data.images.image_corruptor import ImageCorruptor
2020-12-18 16:18:34 +00:00
from trainer.ExtensibleTrainer import ExtensibleTrainer
from utils import util
2020-11-27 19:03:16 +00:00
def image_2_tensor(impath, max_size=None):
img = Image.open(impath)
2020-11-27 19:03:16 +00:00
if max_size is not None:
factor = min(max_size / img.width, max_size / img.height)
new_size = (int(math.ceil(img.width * factor)), int(math.ceil(img.height * factor)))
2020-11-27 19:03:16 +00:00
img = img.resize(new_size, Image.LANCZOS)
2020-11-27 19:03:16 +00:00
'''
# Useful for setting an image to an exact size.
h_gap = img.height - desired_size[1]
w_gap = img.width - desired_size[0]
assert h_gap >= 0 and w_gap >= 0
ht = h_gap // 2
hb = desired_size[1] + ht
wl = w_gap // 2
wr = desired_size[1] + wl
2020-11-27 19:03:16 +00:00
'''
timg = torchvision.transforms.ToTensor()(img).unsqueeze(0)
2020-11-27 19:03:16 +00:00
#if desired_size is not None:
# timg = timg[:, :3, ht:hb, wl:wr]
# assert timg.shape[2] == desired_size[1] and timg.shape[3] == desired_size[0]
#else:
# Enforce that the input must have a input dimension that is a factor of 16.
b, c, h, w = timg.shape
h = (h // 16) * 16
w = (w // 16) * 16
timg = timg[:, :3, :h, :w]
return timg
def interpolate_lr(hr, scale):
return F.interpolate(hr, scale_factor=1 / scale, mode="area")
def fetch_latents_for_image(gen, img, scale, lr_infer=interpolate_lr):
z, _, _ = gen(gt=img,
lr=lr_infer(img, scale),
epses=[],
reverse=False,
add_gt_noise=False)
return z
def fetch_latents_for_images(gen, imgs, scale, lr_infer=interpolate_lr):
latents = []
for img in imgs:
z, _, _ = gen(gt=img,
lr=lr_infer(img, scale),
epses=[],
reverse=False,
add_gt_noise=False)
latents.append(z)
return latents
def fetch_spatial_metrics_for_latents(latents):
dt_scales = []
dt_biases = []
for i in range(len(latents)):
latent = torch.stack(latents[i], dim=-1).squeeze(0)
s = latent.std(dim=[1, 2, 3]).view(1,-1,1,1)
b = latent.mean(dim=[1, 2, 3]).view(1,-1,1,1)
dt_scales.append(s)
dt_biases.append(b)
return dt_scales, dt_biases
2020-11-27 19:03:16 +00:00
def spatial_norm(latents, exclusion_list=[]):
nlatents = []
for i in range(len(latents)):
latent = latents[i]
2020-11-27 19:03:16 +00:00
if i in exclusion_list:
nlatents.append(latent)
else:
b, c, h, w = latent.shape
s = latent.std(dim=[2, 3]).view(1,c,1,1)
b = latent.mean(dim=[2, 3]).view(1,c,1,1)
nlatents.append((latents[i] - b) / s)
return nlatents
2020-11-27 19:03:16 +00:00
def local_norm(latents, exclusion_list=[]):
nlatents = []
for i in range(len(latents)):
latent = latents[i]
2020-11-27 19:03:16 +00:00
if i in exclusion_list:
nlatents.append(latent)
else:
b, c, h, w = latent.shape
s = latent.std(dim=[1]).view(1,1,h,w)
b = latent.mean(dim=[1]).view(1,1,h,w)
nlatents.append((latents[i] - b) / s)
return nlatents
2020-11-27 19:03:16 +00:00
# Extracts a rectangle of the same shape as <lat> from <ref> and returns it. This is taken from the center of <ref>
def extract_center_latent(ref, lat):
_, _, h, w = lat.shape
_, _, rh, rw = ref.shape
dw = (rw - w) / 2
dh = (rh - h) / 2
return ref[:, :, math.floor(dh):-math.ceil(dh), math.floor(dw):-math.ceil(dw)]
2020-11-29 16:33:29 +00:00
def linear_interpolation(latents1, latents2, proportion):
return [l1*proportion+l2*(1-proportion) for l1, l2 in zip(latents1, latents2)]
def slerp(latents1, latents2, proportion):
res = []
for low, high in zip(latents1, latents2):
low_norm = low / torch.norm(low, dim=[2,3], keepdim=True)
high_norm = high / torch.norm(high, dim=[2,3], keepdim=True)
omega = torch.acos((low_norm * high_norm).sum(1))
so = torch.sin(omega)
res.append((torch.sin((1.0 - proportion) * omega) / so).unsqueeze(1) * low + (torch.sin(proportion * omega) / so).unsqueeze(1) * high)
return res
2020-12-15 06:59:49 +00:00
def create_interpolation_video(gen, lq, output_file, latents1, latents2, steps=10, prefix=''):
# Outputs a series of images interpolated from [latents1] to [latents2]. image 0 biases towards latents2.
2020-11-29 16:33:29 +00:00
for i in range(steps):
proportion = i / (steps-1)
lats = linear_interpolation(latents1, latents2, proportion)
2020-12-22 22:42:38 +00:00
hr, _, _ = gen(lr=lq,
2020-11-29 16:33:29 +00:00
z=lats[0],
reverse=True,
epses=lats,
add_gt_noise=False)
torchvision.transforms.ToPILImage()(hr.squeeze(0).cpu())
2020-12-15 06:59:49 +00:00
torchvision.utils.save_image(hr.cpu(), os.path.join(output_file, "%s_%i.png" % (prefix, i)))
2020-11-29 16:33:29 +00:00
# Stopped using this because PILs animated gif output is total crap.
#images[0].save(output_file, save_all=True, append_images=images[1:], duration=80, loop=0)
if __name__ == "__main__":
#### options
torch.backends.cudnn.benchmark = True
srg_analyze = False
parser = argparse.ArgumentParser()
2020-12-30 03:24:41 +00:00
parser.add_argument('-opt', type=str, help='Path to options YAML file.', default='../options/train_exd_imgsetext_srflow_bigboi_ganbase.yml')
#parser.add_argument('-opt', type=str, help='Path to options YAML file.', default='../../options/train_exd_imgsetext_srflow_bigboi_frompsnr.yml')
opt = option.parse(parser.parse_args().opt, is_train=False)
opt = option.dict_to_nonedict(opt)
utils.util.loaded_options = opt
util.mkdirs(
(path for key, path in opt['path'].items()
if not key == 'experiments_root' and 'pretrain_model' not in key and 'resume' not in key))
util.setup_logger('base', opt['path']['log'], 'test_' + opt['name'], level=logging.INFO,
screen=True, tofile=True)
logger = logging.getLogger('base')
logger.info(option.dict2str(opt))
model = ExtensibleTrainer(opt)
gen = model.networks['generator']
gen.eval()
2020-12-22 22:42:38 +00:00
mode = "feed_through" # temperature | restore | latent_transfer | feed_through
2020-12-30 03:24:41 +00:00
#imgs_to_resample_pattern = "F:\\4k6k\\datasets\\ns_images\\adrianna\\val2\\lr\\*"
imgs_to_resample_pattern = "F:\\4k6k\\datasets\\ns_images\\adrianna\\analyze\\analyze_xx\\*"
#imgs_to_resample_pattern = "F:\\4k6k\\datasets\\ns_images\\imagesets\\images-half\\*lanette*"
2020-11-27 19:03:16 +00:00
scale = 2
2020-12-15 06:59:49 +00:00
resample_factor = 2 # When != 1, the HR image is upsampled by this factor using a bicubic to get the local latents. E.g. set this to '2' to get 2x upsampling.
2020-11-29 16:33:29 +00:00
temperature = 1
2020-12-15 06:59:49 +00:00
output_path = "..\\..\\results\\latent_playground"
# Data types <- used to perform latent transfer.
data_path = "F:\\4k6k\\datasets\\ns_images\\imagesets\\images-half"
2020-12-15 06:59:49 +00:00
#data_type_filters = ["*alexa*", "*lanette*", "*80755*", "*joli_high*"]
data_type_filters = ["*lanette*"]
max_size = 1600 # Should be set to 2x the largest single dimension of the input space, otherwise an error will occur.
max_ref_datatypes = 30 # Only picks this many images from the above data types to sample from.
interpolation_steps = 30
with torch.no_grad():
# Compute latent variables for the reference images.
if mode == "latent_transfer":
# Just get the **one** result for each pattern and use that latent.
dt_imgs = [glob(os.path.join(data_path, p))[-5] for p in data_type_filters]
2020-11-27 19:03:16 +00:00
dt_transfers = [image_2_tensor(i, max_size) for i in dt_imgs]
# Downsample the images because they are often just too big to feed through the network (probably needs to be parameterized)
2020-12-15 06:59:49 +00:00
#for j in range(len(dt_transfers)):
# if min(dt_transfers[j].shape[2], dt_transfers[j].shape[3]) > 1600:
# dt_transfers[j] = F.interpolate(dt_transfers[j], scale_factor=1 / 2, mode='area')
corruptor = ImageCorruptor({'fixed_corruptions': ['jpeg-medium', 'gaussian_blur_3']})
def corrupt_and_downsample(img, scale):
img = F.interpolate(img, scale_factor=1 / scale, mode="area")
from data.util import torch2cv, cv2torch
cvimg = torch2cv(img)
cvimg = corruptor.corrupt_images([cvimg])[0]
img = cv2torch(cvimg)
return img
dt_latents = [fetch_latents_for_image(gen, i, scale, corrupt_and_downsample) for i in dt_transfers]
# Fetch the images to resample.
img_files = glob(imgs_to_resample_pattern)
2020-12-15 06:59:49 +00:00
#random.shuffle(img_files)
for im_it, img_file in enumerate(tqdm(img_files)):
2020-11-27 19:03:16 +00:00
t = image_2_tensor(img_file).to(model.env['device'])
2020-12-15 06:59:49 +00:00
if resample_factor > 1:
t = F.interpolate(t, scale_factor=resample_factor, mode="bicubic")
2020-12-15 06:59:49 +00:00
elif resample_factor < 1:
t = F.interpolate(t, scale_factor=resample_factor, mode="area")
# Ensure the input image is a factor of 16.
_, _, h, w = t.shape
h = 16 * (h // 16)
w = 16 * (w // 16)
t = t[:, :, :h, :w]
resample_img = t
# Fetch the latent metrics & latents for each image we are resampling.
latents = fetch_latents_for_images(gen, [resample_img], scale)[0]
multiple_latents = False
if mode == "restore":
latents = local_norm(spatial_norm(latents))
2020-11-27 19:03:16 +00:00
#latents = spatial_norm(latents)
latents = [l * temperature for l in latents]
elif mode == "feed_through":
latents = [torch.randn_like(l) * temperature for l in latents]
elif mode == "latent_transfer":
dts = []
for slat in dt_latents:
assert slat[0].shape[2] >= latents[0].shape[2]
assert slat[0].shape[3] >= latents[0].shape[3]
2020-11-27 19:03:16 +00:00
dts.append([extract_center_latent(sl, l) * temperature for l, sl in zip(latents, slat)])
latents = dts
multiple_latents = True
2020-11-29 16:33:29 +00:00
elif mode == "temperature":
latents = [l * temperature for l in latents]
# Re-compute each image with the new metrics
if not multiple_latents:
lats = [latents]
else:
lats = latents
for j in range(len(lats)):
2020-11-29 16:33:29 +00:00
path = os.path.join(output_path, "%i_%i" % (im_it, j))
os.makedirs(path, exist_ok=True)
2020-12-15 06:59:49 +00:00
torchvision.utils.save_image(resample_img, os.path.join(path, "orig_%i.jpg" % (im_it)))
2020-11-29 16:33:29 +00:00
create_interpolation_video(gen, F.interpolate(resample_img, scale_factor=1/scale, mode="area"),
2020-12-15 06:59:49 +00:00
path, [torch.zeros_like(l) for l in lats[j]], lats[j], prefix=mode)