InvokeAI/ldm/models/diffusion/ddpm.py

2190 lines
76 KiB
Python
Raw Normal View History

2021-12-21 02:23:41 +00:00
"""
wild mixture of
https://github.com/lucidrains/denoising-diffusion-pytorch/blob/7706bdfc6f527f58d33f84b7b522e61e6e3164b3/denoising_diffusion_pytorch/denoising_diffusion_pytorch.py
https://github.com/openai/improved-diffusion/blob/e94489283bb876ac1477d5dd7709bbbd2d9902ce/improved_diffusion/gaussian_diffusion.py
https://github.com/CompVis/taming-transformers
-- merci
"""
import torch
2022-08-23 22:26:28 +00:00
2021-12-21 02:23:41 +00:00
import torch.nn as nn
2022-08-23 22:26:28 +00:00
import os
2021-12-21 02:23:41 +00:00
import numpy as np
import pytorch_lightning as pl
from torch.optim.lr_scheduler import LambdaLR
from einops import rearrange, repeat
from contextlib import contextmanager
from functools import partial
from tqdm import tqdm
from torchvision.utils import make_grid
from pytorch_lightning.utilities.distributed import rank_zero_only
import urllib
2021-12-21 02:23:41 +00:00
from ldm.util import (
log_txt_as_img,
exists,
default,
ismap,
isimage,
mean_flat,
count_params,
instantiate_from_config,
)
2021-12-21 02:23:41 +00:00
from ldm.modules.ema import LitEma
from ldm.modules.distributions.distributions import (
normal_kl,
DiagonalGaussianDistribution,
)
from ldm.models.autoencoder import (
VQModelInterface,
IdentityFirstStage,
AutoencoderKL,
)
from ldm.modules.diffusionmodules.util import (
make_beta_schedule,
extract_into_tensor,
noise_like,
)
from ldm.models.diffusion.ddim import DDIMSampler
2021-12-21 02:23:41 +00:00
__conditioning_keys__ = {
'concat': 'c_concat',
'crossattn': 'c_crossattn',
'adm': 'y',
}
2021-12-21 02:23:41 +00:00
def disabled_train(self, mode=True):
"""Overwrite model.train with this function to make sure train/eval mode
does not change anymore."""
return self
def uniform_on_device(r1, r2, shape, device):
return (r1 - r2) * torch.rand(*shape, device=device) + r2
class DDPM(pl.LightningModule):
# classic DDPM with Gaussian diffusion, in image space
def __init__(
self,
unet_config,
timesteps=1000,
beta_schedule='linear',
loss_type='l2',
ckpt_path=None,
ignore_keys=[],
load_only_unet=False,
monitor='val/loss',
use_ema=True,
first_stage_key='image',
image_size=256,
channels=3,
log_every_t=100,
clip_denoised=True,
linear_start=1e-4,
linear_end=2e-2,
cosine_s=8e-3,
given_betas=None,
original_elbo_weight=0.0,
embedding_reg_weight=0.0,
v_posterior=0.0, # weight for choosing posterior variance as sigma = (1-v) * beta_tilde + v * beta
l_simple_weight=1.0,
conditioning_key=None,
parameterization='eps', # all assuming fixed variance schedules
scheduler_config=None,
use_positional_encodings=False,
learn_logvar=False,
logvar_init=0.0,
):
2021-12-21 02:23:41 +00:00
super().__init__()
assert parameterization in [
'eps',
'x0',
], 'currently only supporting "eps" and "x0"'
2021-12-21 02:23:41 +00:00
self.parameterization = parameterization
print(
f'{self.__class__.__name__}: Running in {self.parameterization}-prediction mode'
)
2021-12-21 02:23:41 +00:00
self.cond_stage_model = None
self.clip_denoised = clip_denoised
self.log_every_t = log_every_t
self.first_stage_key = first_stage_key
self.image_size = image_size # try conv?
self.channels = channels
self.use_positional_encodings = use_positional_encodings
self.model = DiffusionWrapper(unet_config, conditioning_key)
count_params(self.model, verbose=True)
self.use_ema = use_ema
if self.use_ema:
self.model_ema = LitEma(self.model)
print(f'Keeping EMAs of {len(list(self.model_ema.buffers()))}.')
2021-12-21 02:23:41 +00:00
self.use_scheduler = scheduler_config is not None
if self.use_scheduler:
self.scheduler_config = scheduler_config
self.v_posterior = v_posterior
self.original_elbo_weight = original_elbo_weight
self.l_simple_weight = l_simple_weight
2022-08-23 22:26:28 +00:00
self.embedding_reg_weight = embedding_reg_weight
2021-12-21 02:23:41 +00:00
if monitor is not None:
self.monitor = monitor
if ckpt_path is not None:
self.init_from_ckpt(
ckpt_path, ignore_keys=ignore_keys, only_model=load_only_unet
)
self.register_schedule(
given_betas=given_betas,
beta_schedule=beta_schedule,
timesteps=timesteps,
linear_start=linear_start,
linear_end=linear_end,
cosine_s=cosine_s,
)
2021-12-21 02:23:41 +00:00
self.loss_type = loss_type
self.learn_logvar = learn_logvar
self.logvar = torch.full(
fill_value=logvar_init, size=(self.num_timesteps,)
)
2021-12-21 02:23:41 +00:00
if self.learn_logvar:
self.logvar = nn.Parameter(self.logvar, requires_grad=True)
def register_schedule(
self,
given_betas=None,
beta_schedule='linear',
timesteps=1000,
linear_start=1e-4,
linear_end=2e-2,
cosine_s=8e-3,
):
2021-12-21 02:23:41 +00:00
if exists(given_betas):
betas = given_betas
else:
betas = make_beta_schedule(
beta_schedule,
timesteps,
linear_start=linear_start,
linear_end=linear_end,
cosine_s=cosine_s,
)
alphas = 1.0 - betas
2021-12-21 02:23:41 +00:00
alphas_cumprod = np.cumprod(alphas, axis=0)
alphas_cumprod_prev = np.append(1.0, alphas_cumprod[:-1])
2021-12-21 02:23:41 +00:00
(timesteps,) = betas.shape
2021-12-21 02:23:41 +00:00
self.num_timesteps = int(timesteps)
self.linear_start = linear_start
self.linear_end = linear_end
assert (
alphas_cumprod.shape[0] == self.num_timesteps
), 'alphas have to be defined for each timestep'
2021-12-21 02:23:41 +00:00
to_torch = partial(torch.tensor, dtype=torch.float32)
self.register_buffer('betas', to_torch(betas))
self.register_buffer('alphas_cumprod', to_torch(alphas_cumprod))
self.register_buffer(
'alphas_cumprod_prev', to_torch(alphas_cumprod_prev)
)
2021-12-21 02:23:41 +00:00
# calculations for diffusion q(x_t | x_{t-1}) and others
self.register_buffer(
'sqrt_alphas_cumprod', to_torch(np.sqrt(alphas_cumprod))
)
self.register_buffer(
'sqrt_one_minus_alphas_cumprod',
to_torch(np.sqrt(1.0 - alphas_cumprod)),
)
self.register_buffer(
'log_one_minus_alphas_cumprod',
to_torch(np.log(1.0 - alphas_cumprod)),
)
self.register_buffer(
'sqrt_recip_alphas_cumprod',
to_torch(np.sqrt(1.0 / alphas_cumprod)),
)
self.register_buffer(
'sqrt_recipm1_alphas_cumprod',
to_torch(np.sqrt(1.0 / alphas_cumprod - 1)),
)
2021-12-21 02:23:41 +00:00
# calculations for posterior q(x_{t-1} | x_t, x_0)
posterior_variance = (1 - self.v_posterior) * betas * (
1.0 - alphas_cumprod_prev
) / (1.0 - alphas_cumprod) + self.v_posterior * betas
2021-12-21 02:23:41 +00:00
# above: equal to 1. / (1. / (1. - alpha_cumprod_tm1) + alpha_t / beta_t)
self.register_buffer(
'posterior_variance', to_torch(posterior_variance)
)
2021-12-21 02:23:41 +00:00
# below: log calculation clipped because the posterior variance is 0 at the beginning of the diffusion chain
self.register_buffer(
'posterior_log_variance_clipped',
to_torch(np.log(np.maximum(posterior_variance, 1e-20))),
)
self.register_buffer(
'posterior_mean_coef1',
to_torch(
betas * np.sqrt(alphas_cumprod_prev) / (1.0 - alphas_cumprod)
),
)
self.register_buffer(
'posterior_mean_coef2',
to_torch(
(1.0 - alphas_cumprod_prev)
* np.sqrt(alphas)
/ (1.0 - alphas_cumprod)
),
)
if self.parameterization == 'eps':
lvlb_weights = self.betas**2 / (
2
* self.posterior_variance
* to_torch(alphas)
* (1 - self.alphas_cumprod)
)
elif self.parameterization == 'x0':
lvlb_weights = (
0.5
* np.sqrt(torch.Tensor(alphas_cumprod))
/ (2.0 * 1 - torch.Tensor(alphas_cumprod))
)
2021-12-21 02:23:41 +00:00
else:
raise NotImplementedError('mu not supported')
2021-12-21 02:23:41 +00:00
# TODO how to choose this term
lvlb_weights[0] = lvlb_weights[1]
self.register_buffer('lvlb_weights', lvlb_weights, persistent=False)
assert not torch.isnan(self.lvlb_weights).all()
@contextmanager
def ema_scope(self, context=None):
if self.use_ema:
self.model_ema.store(self.model.parameters())
self.model_ema.copy_to(self.model)
if context is not None:
print(f'{context}: Switched to EMA weights')
2021-12-21 02:23:41 +00:00
try:
yield None
finally:
if self.use_ema:
self.model_ema.restore(self.model.parameters())
if context is not None:
print(f'{context}: Restored training weights')
2021-12-21 02:23:41 +00:00
def init_from_ckpt(self, path, ignore_keys=list(), only_model=False):
sd = torch.load(path, map_location='cpu')
if 'state_dict' in list(sd.keys()):
sd = sd['state_dict']
2021-12-21 02:23:41 +00:00
keys = list(sd.keys())
for k in keys:
for ik in ignore_keys:
if k.startswith(ik):
print('Deleting key {} from state_dict.'.format(k))
2021-12-21 02:23:41 +00:00
del sd[k]
missing, unexpected = (
self.load_state_dict(sd, strict=False)
if not only_model
else self.model.load_state_dict(sd, strict=False)
)
print(
f'Restored from {path} with {len(missing)} missing and {len(unexpected)} unexpected keys'
)
2021-12-21 02:23:41 +00:00
if len(missing) > 0:
print(f'Missing Keys: {missing}')
2021-12-21 02:23:41 +00:00
if len(unexpected) > 0:
print(f'Unexpected Keys: {unexpected}')
2021-12-21 02:23:41 +00:00
def q_mean_variance(self, x_start, t):
"""
Get the distribution q(x_t | x_0).
:param x_start: the [N x C x ...] tensor of noiseless inputs.
:param t: the number of diffusion steps (minus 1). Here, 0 means one step.
:return: A tuple (mean, variance, log_variance), all of x_start's shape.
"""
mean = (
extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape)
* x_start
)
variance = extract_into_tensor(
1.0 - self.alphas_cumprod, t, x_start.shape
)
log_variance = extract_into_tensor(
self.log_one_minus_alphas_cumprod, t, x_start.shape
)
2021-12-21 02:23:41 +00:00
return mean, variance, log_variance
def predict_start_from_noise(self, x_t, t, noise):
return (
extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape)
* x_t
- extract_into_tensor(
self.sqrt_recipm1_alphas_cumprod, t, x_t.shape
)
* noise
2021-12-21 02:23:41 +00:00
)
def q_posterior(self, x_start, x_t, t):
posterior_mean = (
extract_into_tensor(self.posterior_mean_coef1, t, x_t.shape)
* x_start
+ extract_into_tensor(self.posterior_mean_coef2, t, x_t.shape)
* x_t
)
posterior_variance = extract_into_tensor(
self.posterior_variance, t, x_t.shape
)
posterior_log_variance_clipped = extract_into_tensor(
self.posterior_log_variance_clipped, t, x_t.shape
)
return (
posterior_mean,
posterior_variance,
posterior_log_variance_clipped,
2021-12-21 02:23:41 +00:00
)
def p_mean_variance(self, x, t, clip_denoised: bool):
model_out = self.model(x, t)
if self.parameterization == 'eps':
2021-12-21 02:23:41 +00:00
x_recon = self.predict_start_from_noise(x, t=t, noise=model_out)
elif self.parameterization == 'x0':
2021-12-21 02:23:41 +00:00
x_recon = model_out
if clip_denoised:
x_recon.clamp_(-1.0, 1.0)
2021-12-21 02:23:41 +00:00
(
model_mean,
posterior_variance,
posterior_log_variance,
) = self.q_posterior(x_start=x_recon, x_t=x, t=t)
2021-12-21 02:23:41 +00:00
return model_mean, posterior_variance, posterior_log_variance
@torch.no_grad()
def p_sample(self, x, t, clip_denoised=True, repeat_noise=False):
b, *_, device = *x.shape, x.device
model_mean, _, model_log_variance = self.p_mean_variance(
x=x, t=t, clip_denoised=clip_denoised
)
2021-12-21 02:23:41 +00:00
noise = noise_like(x.shape, device, repeat_noise)
# no noise when t == 0
nonzero_mask = (1 - (t == 0).float()).reshape(
b, *((1,) * (len(x.shape) - 1))
)
return (
model_mean
+ nonzero_mask * (0.5 * model_log_variance).exp() * noise
)
2021-12-21 02:23:41 +00:00
@torch.no_grad()
def p_sample_loop(self, shape, return_intermediates=False):
device = self.betas.device
b = shape[0]
img = torch.randn(shape, device=device)
intermediates = [img]
for i in tqdm(
reversed(range(0, self.num_timesteps)),
desc='Sampling t',
total=self.num_timesteps,
dynamic_ncols=True,
):
img = self.p_sample(
img,
torch.full((b,), i, device=device, dtype=torch.long),
clip_denoised=self.clip_denoised,
)
2021-12-21 02:23:41 +00:00
if i % self.log_every_t == 0 or i == self.num_timesteps - 1:
intermediates.append(img)
if return_intermediates:
return img, intermediates
return img
@torch.no_grad()
def sample(self, batch_size=16, return_intermediates=False):
image_size = self.image_size
channels = self.channels
return self.p_sample_loop(
(batch_size, channels, image_size, image_size),
return_intermediates=return_intermediates,
)
2021-12-21 02:23:41 +00:00
def q_sample(self, x_start, t, noise=None):
noise = default(noise, lambda: torch.randn_like(x_start))
return (
extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape)
* x_start
+ extract_into_tensor(
self.sqrt_one_minus_alphas_cumprod, t, x_start.shape
)
* noise
)
2021-12-21 02:23:41 +00:00
def get_loss(self, pred, target, mean=True):
if self.loss_type == 'l1':
loss = (target - pred).abs()
if mean:
loss = loss.mean()
elif self.loss_type == 'l2':
if mean:
loss = torch.nn.functional.mse_loss(target, pred)
else:
loss = torch.nn.functional.mse_loss(
target, pred, reduction='none'
)
2021-12-21 02:23:41 +00:00
else:
raise NotImplementedError("unknown loss type '{loss_type}'")
return loss
def p_losses(self, x_start, t, noise=None):
noise = default(noise, lambda: torch.randn_like(x_start))
x_noisy = self.q_sample(x_start=x_start, t=t, noise=noise)
model_out = self.model(x_noisy, t)
loss_dict = {}
if self.parameterization == 'eps':
2021-12-21 02:23:41 +00:00
target = noise
elif self.parameterization == 'x0':
2021-12-21 02:23:41 +00:00
target = x_start
else:
raise NotImplementedError(
f'Paramterization {self.parameterization} not yet supported'
)
2021-12-21 02:23:41 +00:00
loss = self.get_loss(model_out, target, mean=False).mean(dim=[1, 2, 3])
log_prefix = 'train' if self.training else 'val'
loss_dict.update({f'{log_prefix}/loss_simple': loss.mean()})
loss_simple = loss.mean() * self.l_simple_weight
loss_vlb = (self.lvlb_weights[t] * loss).mean()
loss_dict.update({f'{log_prefix}/loss_vlb': loss_vlb})
loss = loss_simple + self.original_elbo_weight * loss_vlb
loss_dict.update({f'{log_prefix}/loss': loss})
return loss, loss_dict
def forward(self, x, *args, **kwargs):
# b, c, h, w, device, img_size, = *x.shape, x.device, self.image_size
# assert h == img_size and w == img_size, f'height and width of image must be {img_size}'
t = torch.randint(
0, self.num_timesteps, (x.shape[0],), device=self.device
).long()
2021-12-21 02:23:41 +00:00
return self.p_losses(x, t, *args, **kwargs)
def get_input(self, batch, k):
x = batch[k]
if len(x.shape) == 3:
x = x[..., None]
x = rearrange(x, 'b h w c -> b c h w')
x = x.to(memory_format=torch.contiguous_format).float()
return x
def shared_step(self, batch):
x = self.get_input(batch, self.first_stage_key)
loss, loss_dict = self(x)
return loss, loss_dict
def training_step(self, batch, batch_idx):
loss, loss_dict = self.shared_step(batch)
self.log_dict(
loss_dict, prog_bar=True, logger=True, on_step=True, on_epoch=True
)
2021-12-21 02:23:41 +00:00
self.log(
'global_step',
self.global_step,
prog_bar=True,
logger=True,
on_step=True,
on_epoch=False,
)
2021-12-21 02:23:41 +00:00
if self.use_scheduler:
lr = self.optimizers().param_groups[0]['lr']
self.log(
'lr_abs',
lr,
prog_bar=True,
logger=True,
on_step=True,
on_epoch=False,
)
2021-12-21 02:23:41 +00:00
return loss
@torch.no_grad()
def validation_step(self, batch, batch_idx):
_, loss_dict_no_ema = self.shared_step(batch)
with self.ema_scope():
_, loss_dict_ema = self.shared_step(batch)
loss_dict_ema = {
key + '_ema': loss_dict_ema[key] for key in loss_dict_ema
}
self.log_dict(
loss_dict_no_ema,
prog_bar=False,
logger=True,
on_step=False,
on_epoch=True,
)
self.log_dict(
loss_dict_ema,
prog_bar=False,
logger=True,
on_step=False,
on_epoch=True,
)
2021-12-21 02:23:41 +00:00
def on_train_batch_end(self, *args, **kwargs):
if self.use_ema:
self.model_ema(self.model)
def _get_rows_from_list(self, samples):
n_imgs_per_row = len(samples)
denoise_grid = rearrange(samples, 'n b c h w -> b n c h w')
denoise_grid = rearrange(denoise_grid, 'b n c h w -> (b n) c h w')
denoise_grid = make_grid(denoise_grid, nrow=n_imgs_per_row)
return denoise_grid
@torch.no_grad()
def log_images(
self, batch, N=8, n_row=2, sample=True, return_keys=None, **kwargs
):
2021-12-21 02:23:41 +00:00
log = dict()
x = self.get_input(batch, self.first_stage_key)
N = min(x.shape[0], N)
n_row = min(x.shape[0], n_row)
x = x.to(self.device)[:N]
log['inputs'] = x
2021-12-21 02:23:41 +00:00
# get diffusion row
diffusion_row = list()
x_start = x[:n_row]
for t in range(self.num_timesteps):
if t % self.log_every_t == 0 or t == self.num_timesteps - 1:
t = repeat(torch.tensor([t]), '1 -> b', b=n_row)
t = t.to(self.device).long()
noise = torch.randn_like(x_start)
x_noisy = self.q_sample(x_start=x_start, t=t, noise=noise)
diffusion_row.append(x_noisy)
log['diffusion_row'] = self._get_rows_from_list(diffusion_row)
2021-12-21 02:23:41 +00:00
if sample:
# get denoise row
with self.ema_scope('Plotting'):
samples, denoise_row = self.sample(
batch_size=N, return_intermediates=True
)
2021-12-21 02:23:41 +00:00
log['samples'] = samples
log['denoise_row'] = self._get_rows_from_list(denoise_row)
2021-12-21 02:23:41 +00:00
if return_keys:
if np.intersect1d(list(log.keys()), return_keys).shape[0] == 0:
return log
else:
return {key: log[key] for key in return_keys}
return log
def configure_optimizers(self):
lr = self.learning_rate
params = list(self.model.parameters())
if self.learn_logvar:
params = params + [self.logvar]
opt = torch.optim.AdamW(params, lr=lr)
return opt
class LatentDiffusion(DDPM):
"""main class"""
def __init__(
self,
first_stage_config,
cond_stage_config,
personalization_config,
num_timesteps_cond=None,
cond_stage_key='image',
cond_stage_trainable=False,
concat_mode=True,
cond_stage_forward=None,
conditioning_key=None,
scale_factor=1.0,
scale_by_std=False,
*args,
**kwargs,
):
2022-08-23 22:26:28 +00:00
2021-12-21 02:23:41 +00:00
self.num_timesteps_cond = default(num_timesteps_cond, 1)
self.scale_by_std = scale_by_std
assert self.num_timesteps_cond <= kwargs['timesteps']
# for backwards compatibility after implementation of DiffusionWrapper
if conditioning_key is None:
conditioning_key = 'concat' if concat_mode else 'crossattn'
if cond_stage_config == '__is_unconditional__':
conditioning_key = None
ckpt_path = kwargs.pop('ckpt_path', None)
ignore_keys = kwargs.pop('ignore_keys', [])
2021-12-21 02:23:41 +00:00
super().__init__(conditioning_key=conditioning_key, *args, **kwargs)
self.concat_mode = concat_mode
self.cond_stage_trainable = cond_stage_trainable
self.cond_stage_key = cond_stage_key
2022-08-23 22:26:28 +00:00
2021-12-21 02:23:41 +00:00
try:
self.num_downs = (
len(first_stage_config.params.ddconfig.ch_mult) - 1
)
2021-12-21 02:23:41 +00:00
except:
self.num_downs = 0
if not scale_by_std:
self.scale_factor = scale_factor
else:
self.register_buffer('scale_factor', torch.tensor(scale_factor))
self.instantiate_first_stage(first_stage_config)
self.instantiate_cond_stage(cond_stage_config)
2022-08-23 22:26:28 +00:00
2021-12-21 02:23:41 +00:00
self.cond_stage_forward = cond_stage_forward
self.clip_denoised = False
self.bbox_tokenizer = None
2021-12-21 02:23:41 +00:00
self.restarted_from_ckpt = False
if ckpt_path is not None:
self.init_from_ckpt(ckpt_path, ignore_keys)
self.restarted_from_ckpt = True
2022-08-23 22:26:28 +00:00
self.cond_stage_model.train = disabled_train
for param in self.cond_stage_model.parameters():
param.requires_grad = False
self.model.eval()
self.model.train = disabled_train
for param in self.model.parameters():
param.requires_grad = False
self.embedding_manager = self.instantiate_embedding_manager(
personalization_config, self.cond_stage_model
)
2022-08-23 22:26:28 +00:00
self.emb_ckpt_counter = 0
# if self.embedding_manager.is_clip:
# self.cond_stage_model.update_embedding_func(self.embedding_manager)
for param in self.embedding_manager.embedding_parameters():
param.requires_grad = True
def make_cond_schedule(
self,
):
self.cond_ids = torch.full(
size=(self.num_timesteps,),
fill_value=self.num_timesteps - 1,
dtype=torch.long,
)
ids = torch.round(
torch.linspace(0, self.num_timesteps - 1, self.num_timesteps_cond)
).long()
self.cond_ids[: self.num_timesteps_cond] = ids
2021-12-21 02:23:41 +00:00
@rank_zero_only
@torch.no_grad()
def on_train_batch_start(self, batch, batch_idx, dataloader_idx):
# only for very first batch
if (
self.scale_by_std
and self.current_epoch == 0
and self.global_step == 0
and batch_idx == 0
and not self.restarted_from_ckpt
):
assert (
self.scale_factor == 1.0
), 'rather not use custom rescaling and std-rescaling simultaneously'
2021-12-21 02:23:41 +00:00
# set rescale weight to 1./std of encodings
print('### USING STD-RESCALING ###')
2021-12-21 02:23:41 +00:00
x = super().get_input(batch, self.first_stage_key)
x = x.to(self.device)
encoder_posterior = self.encode_first_stage(x)
z = self.get_first_stage_encoding(encoder_posterior).detach()
del self.scale_factor
self.register_buffer('scale_factor', 1.0 / z.flatten().std())
print(f'setting self.scale_factor to {self.scale_factor}')
print('### USING STD-RESCALING ###')
def register_schedule(
self,
given_betas=None,
beta_schedule='linear',
timesteps=1000,
linear_start=1e-4,
linear_end=2e-2,
cosine_s=8e-3,
):
super().register_schedule(
given_betas,
beta_schedule,
timesteps,
linear_start,
linear_end,
cosine_s,
)
2021-12-21 02:23:41 +00:00
self.shorten_cond_schedule = self.num_timesteps_cond > 1
if self.shorten_cond_schedule:
self.make_cond_schedule()
def instantiate_first_stage(self, config):
model = instantiate_from_config(config)
self.first_stage_model = model.eval()
self.first_stage_model.train = disabled_train
for param in self.first_stage_model.parameters():
param.requires_grad = False
def instantiate_cond_stage(self, config):
if not self.cond_stage_trainable:
if config == '__is_first_stage__':
print('Using first stage also as cond stage.')
2021-12-21 02:23:41 +00:00
self.cond_stage_model = self.first_stage_model
elif config == '__is_unconditional__':
print(
f'Training {self.__class__.__name__} as an unconditional model.'
)
2021-12-21 02:23:41 +00:00
self.cond_stage_model = None
# self.be_unconditional = True
else:
model = instantiate_from_config(config)
self.cond_stage_model = model.eval()
self.cond_stage_model.train = disabled_train
for param in self.cond_stage_model.parameters():
param.requires_grad = False
else:
assert config != '__is_first_stage__'
assert config != '__is_unconditional__'
try:
model = instantiate_from_config(config)
except urllib.error.URLError:
raise SystemExit(
"* Couldn't load a dependency. Try running scripts/preload_models.py from an internet-conected machine."
)
2021-12-21 02:23:41 +00:00
self.cond_stage_model = model
2022-08-23 22:26:28 +00:00
def instantiate_embedding_manager(self, config, embedder):
model = instantiate_from_config(config, embedder=embedder)
if config.params.get(
'embedding_manager_ckpt', None
): # do not load if missing OR empty string
2022-08-23 22:26:28 +00:00
model.load(config.params.embedding_manager_ckpt)
2022-08-23 22:26:28 +00:00
return model
2021-12-21 02:23:41 +00:00
def _get_denoise_row_from_list(
self, samples, desc='', force_no_decoder_quantization=False
):
2021-12-21 02:23:41 +00:00
denoise_row = []
for zd in tqdm(samples, desc=desc):
denoise_row.append(
self.decode_first_stage(
zd.to(self.device),
force_not_quantize=force_no_decoder_quantization,
)
)
2021-12-21 02:23:41 +00:00
n_imgs_per_row = len(denoise_row)
denoise_row = torch.stack(denoise_row) # n_log_step, n_row, C, H, W
denoise_grid = rearrange(denoise_row, 'n b c h w -> b n c h w')
denoise_grid = rearrange(denoise_grid, 'b n c h w -> (b n) c h w')
denoise_grid = make_grid(denoise_grid, nrow=n_imgs_per_row)
return denoise_grid
def get_first_stage_encoding(self, encoder_posterior):
if isinstance(encoder_posterior, DiagonalGaussianDistribution):
z = encoder_posterior.sample()
elif isinstance(encoder_posterior, torch.Tensor):
z = encoder_posterior
else:
raise NotImplementedError(
f"encoder_posterior of type '{type(encoder_posterior)}' not yet implemented"
)
2021-12-21 02:23:41 +00:00
return self.scale_factor * z
def get_learned_conditioning(self, c):
if self.cond_stage_forward is None:
if hasattr(self.cond_stage_model, 'encode') and callable(
self.cond_stage_model.encode
):
c = self.cond_stage_model.encode(
c, embedding_manager=self.embedding_manager
)
2021-12-21 02:23:41 +00:00
if isinstance(c, DiagonalGaussianDistribution):
c = c.mode()
else:
c = self.cond_stage_model(c)
else:
assert hasattr(self.cond_stage_model, self.cond_stage_forward)
c = getattr(self.cond_stage_model, self.cond_stage_forward)(c)
return c
def meshgrid(self, h, w):
y = torch.arange(0, h).view(h, 1, 1).repeat(1, w, 1)
x = torch.arange(0, w).view(1, w, 1).repeat(h, 1, 1)
arr = torch.cat([y, x], dim=-1)
return arr
def delta_border(self, h, w):
"""
:param h: height
:param w: width
:return: normalized distance to image border,
wtith min distance = 0 at border and max dist = 0.5 at image center
"""
lower_right_corner = torch.tensor([h - 1, w - 1]).view(1, 1, 2)
arr = self.meshgrid(h, w) / lower_right_corner
dist_left_up = torch.min(arr, dim=-1, keepdims=True)[0]
dist_right_down = torch.min(1 - arr, dim=-1, keepdims=True)[0]
edge_dist = torch.min(
torch.cat([dist_left_up, dist_right_down], dim=-1), dim=-1
)[0]
2021-12-21 02:23:41 +00:00
return edge_dist
def get_weighting(self, h, w, Ly, Lx, device):
weighting = self.delta_border(h, w)
weighting = torch.clip(
weighting,
self.split_input_params['clip_min_weight'],
self.split_input_params['clip_max_weight'],
)
weighting = (
weighting.view(1, h * w, 1).repeat(1, 1, Ly * Lx).to(device)
)
2021-12-21 02:23:41 +00:00
if self.split_input_params['tie_braker']:
2021-12-21 02:23:41 +00:00
L_weighting = self.delta_border(Ly, Lx)
L_weighting = torch.clip(
L_weighting,
self.split_input_params['clip_min_tie_weight'],
self.split_input_params['clip_max_tie_weight'],
)
2021-12-21 02:23:41 +00:00
L_weighting = L_weighting.view(1, 1, Ly * Lx).to(device)
weighting = weighting * L_weighting
return weighting
def get_fold_unfold(
self, x, kernel_size, stride, uf=1, df=1
): # todo load once not every time, shorten code
2021-12-21 02:23:41 +00:00
"""
:param x: img of size (bs, c, h, w)
:return: n img crops of size (n, bs, c, kernel_size[0], kernel_size[1])
"""
bs, nc, h, w = x.shape
# number of crops in image
Ly = (h - kernel_size[0]) // stride[0] + 1
Lx = (w - kernel_size[1]) // stride[1] + 1
if uf == 1 and df == 1:
fold_params = dict(
kernel_size=kernel_size, dilation=1, padding=0, stride=stride
)
2021-12-21 02:23:41 +00:00
unfold = torch.nn.Unfold(**fold_params)
fold = torch.nn.Fold(output_size=x.shape[2:], **fold_params)
weighting = self.get_weighting(
kernel_size[0], kernel_size[1], Ly, Lx, x.device
).to(x.dtype)
normalization = fold(weighting).view(
1, 1, h, w
) # normalizes the overlap
weighting = weighting.view(
(1, 1, kernel_size[0], kernel_size[1], Ly * Lx)
)
2021-12-21 02:23:41 +00:00
elif uf > 1 and df == 1:
fold_params = dict(
kernel_size=kernel_size, dilation=1, padding=0, stride=stride
)
2021-12-21 02:23:41 +00:00
unfold = torch.nn.Unfold(**fold_params)
fold_params2 = dict(
kernel_size=(kernel_size[0] * uf, kernel_size[0] * uf),
dilation=1,
padding=0,
stride=(stride[0] * uf, stride[1] * uf),
)
fold = torch.nn.Fold(
output_size=(x.shape[2] * uf, x.shape[3] * uf), **fold_params2
)
weighting = self.get_weighting(
kernel_size[0] * uf, kernel_size[1] * uf, Ly, Lx, x.device
).to(x.dtype)
normalization = fold(weighting).view(
1, 1, h * uf, w * uf
) # normalizes the overlap
weighting = weighting.view(
(1, 1, kernel_size[0] * uf, kernel_size[1] * uf, Ly * Lx)
)
2021-12-21 02:23:41 +00:00
elif df > 1 and uf == 1:
fold_params = dict(
kernel_size=kernel_size, dilation=1, padding=0, stride=stride
)
2021-12-21 02:23:41 +00:00
unfold = torch.nn.Unfold(**fold_params)
fold_params2 = dict(
kernel_size=(kernel_size[0] // df, kernel_size[0] // df),
dilation=1,
padding=0,
stride=(stride[0] // df, stride[1] // df),
)
fold = torch.nn.Fold(
output_size=(x.shape[2] // df, x.shape[3] // df),
**fold_params2,
)
weighting = self.get_weighting(
kernel_size[0] // df, kernel_size[1] // df, Ly, Lx, x.device
).to(x.dtype)
normalization = fold(weighting).view(
1, 1, h // df, w // df
) # normalizes the overlap
weighting = weighting.view(
(1, 1, kernel_size[0] // df, kernel_size[1] // df, Ly * Lx)
)
2021-12-21 02:23:41 +00:00
else:
raise NotImplementedError
return fold, unfold, normalization, weighting
@torch.no_grad()
def get_input(
self,
batch,
k,
return_first_stage_outputs=False,
force_c_encode=False,
cond_key=None,
return_original_cond=False,
bs=None,
):
2021-12-21 02:23:41 +00:00
x = super().get_input(batch, k)
if bs is not None:
x = x[:bs]
x = x.to(self.device)
encoder_posterior = self.encode_first_stage(x)
z = self.get_first_stage_encoding(encoder_posterior).detach()
if self.model.conditioning_key is not None:
if cond_key is None:
cond_key = self.cond_stage_key
if cond_key != self.first_stage_key:
if cond_key in ['caption', 'coordinates_bbox']:
xc = batch[cond_key]
elif cond_key == 'class_label':
xc = batch
else:
xc = super().get_input(batch, cond_key).to(self.device)
else:
xc = x
if not self.cond_stage_trainable or force_c_encode:
if isinstance(xc, dict) or isinstance(xc, list):
# import pudb; pudb.set_trace()
c = self.get_learned_conditioning(xc)
else:
c = self.get_learned_conditioning(xc.to(self.device))
else:
c = xc
if bs is not None:
c = c[:bs]
if self.use_positional_encodings:
pos_x, pos_y = self.compute_latent_shifts(batch)
ckey = __conditioning_keys__[self.model.conditioning_key]
c = {ckey: c, 'pos_x': pos_x, 'pos_y': pos_y}
else:
c = None
xc = None
if self.use_positional_encodings:
pos_x, pos_y = self.compute_latent_shifts(batch)
c = {'pos_x': pos_x, 'pos_y': pos_y}
out = [z, c]
if return_first_stage_outputs:
xrec = self.decode_first_stage(z)
out.extend([x, xrec])
if return_original_cond:
out.append(xc)
return out
@torch.no_grad()
def decode_first_stage(
self, z, predict_cids=False, force_not_quantize=False
):
2021-12-21 02:23:41 +00:00
if predict_cids:
if z.dim() == 4:
z = torch.argmax(z.exp(), dim=1).long()
z = self.first_stage_model.quantize.get_codebook_entry(
z, shape=None
)
2021-12-21 02:23:41 +00:00
z = rearrange(z, 'b h w c -> b c h w').contiguous()
z = 1.0 / self.scale_factor * z
2021-12-21 02:23:41 +00:00
if hasattr(self, 'split_input_params'):
if self.split_input_params['patch_distributed_vq']:
ks = self.split_input_params['ks'] # eg. (128, 128)
stride = self.split_input_params['stride'] # eg. (64, 64)
uf = self.split_input_params['vqf']
2021-12-21 02:23:41 +00:00
bs, nc, h, w = z.shape
if ks[0] > h or ks[1] > w:
ks = (min(ks[0], h), min(ks[1], w))
print('reducing Kernel')
2021-12-21 02:23:41 +00:00
if stride[0] > h or stride[1] > w:
stride = (min(stride[0], h), min(stride[1], w))
print('reducing stride')
2021-12-21 02:23:41 +00:00
fold, unfold, normalization, weighting = self.get_fold_unfold(
z, ks, stride, uf=uf
)
2021-12-21 02:23:41 +00:00
z = unfold(z) # (bn, nc * prod(**ks), L)
# 1. Reshape to img shape
z = z.view(
(z.shape[0], -1, ks[0], ks[1], z.shape[-1])
) # (bn, nc, ks[0], ks[1], L )
2021-12-21 02:23:41 +00:00
# 2. apply model loop over last dim
if isinstance(self.first_stage_model, VQModelInterface):
output_list = [
self.first_stage_model.decode(
z[:, :, :, :, i],
force_not_quantize=predict_cids
or force_not_quantize,
)
for i in range(z.shape[-1])
]
2021-12-21 02:23:41 +00:00
else:
output_list = [
self.first_stage_model.decode(z[:, :, :, :, i])
for i in range(z.shape[-1])
]
2021-12-21 02:23:41 +00:00
o = torch.stack(
output_list, axis=-1
) # # (bn, nc, ks[0], ks[1], L)
2021-12-21 02:23:41 +00:00
o = o * weighting
# Reverse 1. reshape to img shape
o = o.view(
(o.shape[0], -1, o.shape[-1])
) # (bn, nc * ks[0] * ks[1], L)
2021-12-21 02:23:41 +00:00
# stitch crops together
decoded = fold(o)
decoded = decoded / normalization # norm is shape (1, 1, h, w)
return decoded
else:
if isinstance(self.first_stage_model, VQModelInterface):
return self.first_stage_model.decode(
z,
force_not_quantize=predict_cids or force_not_quantize,
)
2021-12-21 02:23:41 +00:00
else:
return self.first_stage_model.decode(z)
else:
if isinstance(self.first_stage_model, VQModelInterface):
return self.first_stage_model.decode(
z, force_not_quantize=predict_cids or force_not_quantize
)
2021-12-21 02:23:41 +00:00
else:
return self.first_stage_model.decode(z)
# same as above but without decorator
def differentiable_decode_first_stage(
self, z, predict_cids=False, force_not_quantize=False
):
2021-12-21 02:23:41 +00:00
if predict_cids:
if z.dim() == 4:
z = torch.argmax(z.exp(), dim=1).long()
z = self.first_stage_model.quantize.get_codebook_entry(
z, shape=None
)
2021-12-21 02:23:41 +00:00
z = rearrange(z, 'b h w c -> b c h w').contiguous()
z = 1.0 / self.scale_factor * z
2021-12-21 02:23:41 +00:00
if hasattr(self, 'split_input_params'):
if self.split_input_params['patch_distributed_vq']:
ks = self.split_input_params['ks'] # eg. (128, 128)
stride = self.split_input_params['stride'] # eg. (64, 64)
uf = self.split_input_params['vqf']
2021-12-21 02:23:41 +00:00
bs, nc, h, w = z.shape
if ks[0] > h or ks[1] > w:
ks = (min(ks[0], h), min(ks[1], w))
print('reducing Kernel')
2021-12-21 02:23:41 +00:00
if stride[0] > h or stride[1] > w:
stride = (min(stride[0], h), min(stride[1], w))
print('reducing stride')
2021-12-21 02:23:41 +00:00
fold, unfold, normalization, weighting = self.get_fold_unfold(
z, ks, stride, uf=uf
)
2021-12-21 02:23:41 +00:00
z = unfold(z) # (bn, nc * prod(**ks), L)
# 1. Reshape to img shape
z = z.view(
(z.shape[0], -1, ks[0], ks[1], z.shape[-1])
) # (bn, nc, ks[0], ks[1], L )
2021-12-21 02:23:41 +00:00
# 2. apply model loop over last dim
if isinstance(self.first_stage_model, VQModelInterface):
output_list = [
self.first_stage_model.decode(
z[:, :, :, :, i],
force_not_quantize=predict_cids
or force_not_quantize,
)
for i in range(z.shape[-1])
]
2021-12-21 02:23:41 +00:00
else:
output_list = [
self.first_stage_model.decode(z[:, :, :, :, i])
for i in range(z.shape[-1])
]
2021-12-21 02:23:41 +00:00
o = torch.stack(
output_list, axis=-1
) # # (bn, nc, ks[0], ks[1], L)
2021-12-21 02:23:41 +00:00
o = o * weighting
# Reverse 1. reshape to img shape
o = o.view(
(o.shape[0], -1, o.shape[-1])
) # (bn, nc * ks[0] * ks[1], L)
2021-12-21 02:23:41 +00:00
# stitch crops together
decoded = fold(o)
decoded = decoded / normalization # norm is shape (1, 1, h, w)
return decoded
else:
if isinstance(self.first_stage_model, VQModelInterface):
return self.first_stage_model.decode(
z,
force_not_quantize=predict_cids or force_not_quantize,
)
2021-12-21 02:23:41 +00:00
else:
return self.first_stage_model.decode(z)
else:
if isinstance(self.first_stage_model, VQModelInterface):
return self.first_stage_model.decode(
z, force_not_quantize=predict_cids or force_not_quantize
)
2021-12-21 02:23:41 +00:00
else:
return self.first_stage_model.decode(z)
@torch.no_grad()
def encode_first_stage(self, x):
if hasattr(self, 'split_input_params'):
if self.split_input_params['patch_distributed_vq']:
ks = self.split_input_params['ks'] # eg. (128, 128)
stride = self.split_input_params['stride'] # eg. (64, 64)
df = self.split_input_params['vqf']
2021-12-21 02:23:41 +00:00
self.split_input_params['original_image_size'] = x.shape[-2:]
bs, nc, h, w = x.shape
if ks[0] > h or ks[1] > w:
ks = (min(ks[0], h), min(ks[1], w))
print('reducing Kernel')
2021-12-21 02:23:41 +00:00
if stride[0] > h or stride[1] > w:
stride = (min(stride[0], h), min(stride[1], w))
print('reducing stride')
2021-12-21 02:23:41 +00:00
fold, unfold, normalization, weighting = self.get_fold_unfold(
x, ks, stride, df=df
)
2021-12-21 02:23:41 +00:00
z = unfold(x) # (bn, nc * prod(**ks), L)
# Reshape to img shape
z = z.view(
(z.shape[0], -1, ks[0], ks[1], z.shape[-1])
) # (bn, nc, ks[0], ks[1], L )
2021-12-21 02:23:41 +00:00
output_list = [
self.first_stage_model.encode(z[:, :, :, :, i])
for i in range(z.shape[-1])
]
2021-12-21 02:23:41 +00:00
o = torch.stack(output_list, axis=-1)
o = o * weighting
# Reverse reshape to img shape
o = o.view(
(o.shape[0], -1, o.shape[-1])
) # (bn, nc * ks[0] * ks[1], L)
2021-12-21 02:23:41 +00:00
# stitch crops together
decoded = fold(o)
decoded = decoded / normalization
return decoded
else:
return self.first_stage_model.encode(x)
else:
return self.first_stage_model.encode(x)
def shared_step(self, batch, **kwargs):
x, c = self.get_input(batch, self.first_stage_key)
loss = self(x, c)
return loss
def forward(self, x, c, *args, **kwargs):
t = torch.randint(
0, self.num_timesteps, (x.shape[0],), device=self.device
).long()
2021-12-21 02:23:41 +00:00
if self.model.conditioning_key is not None:
assert c is not None
if self.cond_stage_trainable:
c = self.get_learned_conditioning(c)
if self.shorten_cond_schedule: # TODO: drop this option
tc = self.cond_ids[t].to(self.device)
c = self.q_sample(
x_start=c, t=tc, noise=torch.randn_like(c.float())
)
2022-08-23 22:26:28 +00:00
2021-12-21 02:23:41 +00:00
return self.p_losses(x, c, t, *args, **kwargs)
def _rescale_annotations(
self, bboxes, crop_coordinates
): # TODO: move to dataset
2021-12-21 02:23:41 +00:00
def rescale_bbox(bbox):
x0 = clamp((bbox[0] - crop_coordinates[0]) / crop_coordinates[2])
y0 = clamp((bbox[1] - crop_coordinates[1]) / crop_coordinates[3])
w = min(bbox[2] / crop_coordinates[2], 1 - x0)
h = min(bbox[3] / crop_coordinates[3], 1 - y0)
return x0, y0, w, h
return [rescale_bbox(b) for b in bboxes]
def apply_model(self, x_noisy, t, cond, return_ids=False):
if isinstance(cond, dict):
# hybrid case, cond is exptected to be a dict
pass
else:
if not isinstance(cond, list):
cond = [cond]
key = (
'c_concat'
if self.model.conditioning_key == 'concat'
else 'c_crossattn'
)
2021-12-21 02:23:41 +00:00
cond = {key: cond}
if hasattr(self, 'split_input_params'):
assert (
len(cond) == 1
) # todo can only deal with one conditioning atm
assert not return_ids
ks = self.split_input_params['ks'] # eg. (128, 128)
stride = self.split_input_params['stride'] # eg. (64, 64)
2021-12-21 02:23:41 +00:00
h, w = x_noisy.shape[-2:]
fold, unfold, normalization, weighting = self.get_fold_unfold(
x_noisy, ks, stride
)
2021-12-21 02:23:41 +00:00
z = unfold(x_noisy) # (bn, nc * prod(**ks), L)
# Reshape to img shape
z = z.view(
(z.shape[0], -1, ks[0], ks[1], z.shape[-1])
) # (bn, nc, ks[0], ks[1], L )
2021-12-21 02:23:41 +00:00
z_list = [z[:, :, :, :, i] for i in range(z.shape[-1])]
if (
self.cond_stage_key
in ['image', 'LR_image', 'segmentation', 'bbox_img']
and self.model.conditioning_key
): # todo check for completeness
2021-12-21 02:23:41 +00:00
c_key = next(iter(cond.keys())) # get key
c = next(iter(cond.values())) # get value
assert (
len(c) == 1
) # todo extend to list with more than one elem
2021-12-21 02:23:41 +00:00
c = c[0] # get element
c = unfold(c)
c = c.view(
(c.shape[0], -1, ks[0], ks[1], c.shape[-1])
) # (bn, nc, ks[0], ks[1], L )
2021-12-21 02:23:41 +00:00
cond_list = [
{c_key: [c[:, :, :, :, i]]} for i in range(c.shape[-1])
]
2021-12-21 02:23:41 +00:00
elif self.cond_stage_key == 'coordinates_bbox':
assert (
'original_image_size' in self.split_input_params
), 'BoudingBoxRescaling is missing original_image_size'
2021-12-21 02:23:41 +00:00
# assuming padding of unfold is always 0 and its dilation is always 1
n_patches_per_row = int((w - ks[0]) / stride[0] + 1)
full_img_h, full_img_w = self.split_input_params[
'original_image_size'
]
2021-12-21 02:23:41 +00:00
# as we are operating on latents, we need the factor from the original image size to the
# spatial latent size to properly rescale the crops for regenerating the bbox annotations
num_downs = self.first_stage_model.encoder.num_resolutions - 1
rescale_latent = 2 ** (num_downs)
# get top left postions of patches as conforming for the bbbox tokenizer, therefore we
# need to rescale the tl patch coordinates to be in between (0,1)
tl_patch_coordinates = [
(
rescale_latent
* stride[0]
* (patch_nr % n_patches_per_row)
/ full_img_w,
rescale_latent
* stride[1]
* (patch_nr // n_patches_per_row)
/ full_img_h,
)
for patch_nr in range(z.shape[-1])
]
2021-12-21 02:23:41 +00:00
# patch_limits are tl_coord, width and height coordinates as (x_tl, y_tl, h, w)
patch_limits = [
(
x_tl,
y_tl,
rescale_latent * ks[0] / full_img_w,
rescale_latent * ks[1] / full_img_h,
)
for x_tl, y_tl in tl_patch_coordinates
]
2021-12-21 02:23:41 +00:00
# patch_values = [(np.arange(x_tl,min(x_tl+ks, 1.)),np.arange(y_tl,min(y_tl+ks, 1.))) for x_tl, y_tl in tl_patch_coordinates]
# tokenize crop coordinates for the bounding boxes of the respective patches
patch_limits_tknzd = [
torch.LongTensor(self.bbox_tokenizer._crop_encoder(bbox))[
None
].to(self.device)
for bbox in patch_limits
] # list of length l with tensors of shape (1, 2)
2021-12-21 02:23:41 +00:00
print(patch_limits_tknzd[0].shape)
# cut tknzd crop position from conditioning
assert isinstance(
cond, dict
), 'cond must be dict to be fed into model'
2021-12-21 02:23:41 +00:00
cut_cond = cond['c_crossattn'][0][..., :-2].to(self.device)
print(cut_cond.shape)
adapted_cond = torch.stack(
[
torch.cat([cut_cond, p], dim=1)
for p in patch_limits_tknzd
]
)
2021-12-21 02:23:41 +00:00
adapted_cond = rearrange(adapted_cond, 'l b n -> (l b) n')
print(adapted_cond.shape)
adapted_cond = self.get_learned_conditioning(adapted_cond)
print(adapted_cond.shape)
adapted_cond = rearrange(
adapted_cond, '(l b) n d -> l b n d', l=z.shape[-1]
)
2021-12-21 02:23:41 +00:00
print(adapted_cond.shape)
cond_list = [{'c_crossattn': [e]} for e in adapted_cond]
else:
cond_list = [
cond for i in range(z.shape[-1])
] # Todo make this more efficient
2021-12-21 02:23:41 +00:00
# apply model by loop over crops
output_list = [
self.model(z_list[i], t, **cond_list[i])
for i in range(z.shape[-1])
]
assert not isinstance(
output_list[0], tuple
) # todo cant deal with multiple model outputs check this never happens
2021-12-21 02:23:41 +00:00
o = torch.stack(output_list, axis=-1)
o = o * weighting
# Reverse reshape to img shape
o = o.view(
(o.shape[0], -1, o.shape[-1])
) # (bn, nc * ks[0] * ks[1], L)
2021-12-21 02:23:41 +00:00
# stitch crops together
x_recon = fold(o) / normalization
else:
x_recon = self.model(x_noisy, t, **cond)
if isinstance(x_recon, tuple) and not return_ids:
return x_recon[0]
else:
return x_recon
def _predict_eps_from_xstart(self, x_t, t, pred_xstart):
return (
extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape)
* x_t
- pred_xstart
) / extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape)
2021-12-21 02:23:41 +00:00
def _prior_bpd(self, x_start):
"""
Get the prior KL term for the variational lower-bound, measured in
bits-per-dim.
This term can't be optimized, as it only depends on the encoder.
:param x_start: the [N x C x ...] tensor of inputs.
:return: a batch of [N] KL values (in bits), one per batch element.
"""
batch_size = x_start.shape[0]
t = torch.tensor(
[self.num_timesteps - 1] * batch_size, device=x_start.device
)
2021-12-21 02:23:41 +00:00
qt_mean, _, qt_log_variance = self.q_mean_variance(x_start, t)
kl_prior = normal_kl(
mean1=qt_mean, logvar1=qt_log_variance, mean2=0.0, logvar2=0.0
)
2021-12-21 02:23:41 +00:00
return mean_flat(kl_prior) / np.log(2.0)
def p_losses(self, x_start, cond, t, noise=None):
noise = default(noise, lambda: torch.randn_like(x_start))
x_noisy = self.q_sample(x_start=x_start, t=t, noise=noise)
model_output = self.apply_model(x_noisy, t, cond)
loss_dict = {}
prefix = 'train' if self.training else 'val'
if self.parameterization == 'x0':
2021-12-21 02:23:41 +00:00
target = x_start
elif self.parameterization == 'eps':
2021-12-21 02:23:41 +00:00
target = noise
else:
raise NotImplementedError()
loss_simple = self.get_loss(model_output, target, mean=False).mean(
[1, 2, 3]
)
2021-12-21 02:23:41 +00:00
loss_dict.update({f'{prefix}/loss_simple': loss_simple.mean()})
logvar_t = self.logvar[t].to(self.device)
loss = loss_simple / torch.exp(logvar_t) + logvar_t
# loss = loss_simple / torch.exp(self.logvar) + self.logvar
if self.learn_logvar:
loss_dict.update({f'{prefix}/loss_gamma': loss.mean()})
loss_dict.update({'logvar': self.logvar.data.mean()})
loss = self.l_simple_weight * loss.mean()
loss_vlb = self.get_loss(model_output, target, mean=False).mean(
dim=(1, 2, 3)
)
2021-12-21 02:23:41 +00:00
loss_vlb = (self.lvlb_weights[t] * loss_vlb).mean()
loss_dict.update({f'{prefix}/loss_vlb': loss_vlb})
loss += self.original_elbo_weight * loss_vlb
2021-12-21 02:23:41 +00:00
loss_dict.update({f'{prefix}/loss': loss})
2022-08-23 22:26:28 +00:00
if self.embedding_reg_weight > 0:
loss_embedding_reg = (
self.embedding_manager.embedding_to_coarse_loss().mean()
)
2022-08-23 22:26:28 +00:00
loss_dict.update({f'{prefix}/loss_emb_reg': loss_embedding_reg})
loss += self.embedding_reg_weight * loss_embedding_reg
2022-08-23 22:26:28 +00:00
loss_dict.update({f'{prefix}/loss': loss})
2021-12-21 02:23:41 +00:00
return loss, loss_dict
def p_mean_variance(
self,
x,
c,
t,
clip_denoised: bool,
return_codebook_ids=False,
quantize_denoised=False,
return_x0=False,
score_corrector=None,
corrector_kwargs=None,
):
2021-12-21 02:23:41 +00:00
t_in = t
model_out = self.apply_model(
x, t_in, c, return_ids=return_codebook_ids
)
2021-12-21 02:23:41 +00:00
if score_corrector is not None:
assert self.parameterization == 'eps'
model_out = score_corrector.modify_score(
self, model_out, x, t, c, **corrector_kwargs
)
2021-12-21 02:23:41 +00:00
if return_codebook_ids:
model_out, logits = model_out
if self.parameterization == 'eps':
2021-12-21 02:23:41 +00:00
x_recon = self.predict_start_from_noise(x, t=t, noise=model_out)
elif self.parameterization == 'x0':
2021-12-21 02:23:41 +00:00
x_recon = model_out
else:
raise NotImplementedError()
if clip_denoised:
x_recon.clamp_(-1.0, 1.0)
2021-12-21 02:23:41 +00:00
if quantize_denoised:
x_recon, _, [_, _, indices] = self.first_stage_model.quantize(
x_recon
)
(
model_mean,
posterior_variance,
posterior_log_variance,
) = self.q_posterior(x_start=x_recon, x_t=x, t=t)
2021-12-21 02:23:41 +00:00
if return_codebook_ids:
return (
model_mean,
posterior_variance,
posterior_log_variance,
logits,
)
2021-12-21 02:23:41 +00:00
elif return_x0:
return (
model_mean,
posterior_variance,
posterior_log_variance,
x_recon,
)
2021-12-21 02:23:41 +00:00
else:
return model_mean, posterior_variance, posterior_log_variance
@torch.no_grad()
def p_sample(
self,
x,
c,
t,
clip_denoised=False,
repeat_noise=False,
return_codebook_ids=False,
quantize_denoised=False,
return_x0=False,
temperature=1.0,
noise_dropout=0.0,
score_corrector=None,
corrector_kwargs=None,
):
2021-12-21 02:23:41 +00:00
b, *_, device = *x.shape, x.device
outputs = self.p_mean_variance(
x=x,
c=c,
t=t,
clip_denoised=clip_denoised,
return_codebook_ids=return_codebook_ids,
quantize_denoised=quantize_denoised,
return_x0=return_x0,
score_corrector=score_corrector,
corrector_kwargs=corrector_kwargs,
)
2021-12-21 02:23:41 +00:00
if return_codebook_ids:
raise DeprecationWarning('Support dropped.')
2021-12-21 02:23:41 +00:00
model_mean, _, model_log_variance, logits = outputs
elif return_x0:
model_mean, _, model_log_variance, x0 = outputs
else:
model_mean, _, model_log_variance = outputs
noise = noise_like(x.shape, device, repeat_noise) * temperature
if noise_dropout > 0.0:
2021-12-21 02:23:41 +00:00
noise = torch.nn.functional.dropout(noise, p=noise_dropout)
# no noise when t == 0
nonzero_mask = (1 - (t == 0).float()).reshape(
b, *((1,) * (len(x.shape) - 1))
)
2021-12-21 02:23:41 +00:00
if return_codebook_ids:
return model_mean + nonzero_mask * (
0.5 * model_log_variance
).exp() * noise, logits.argmax(dim=1)
2021-12-21 02:23:41 +00:00
if return_x0:
return (
model_mean
+ nonzero_mask * (0.5 * model_log_variance).exp() * noise,
x0,
)
2021-12-21 02:23:41 +00:00
else:
return (
model_mean
+ nonzero_mask * (0.5 * model_log_variance).exp() * noise
)
2021-12-21 02:23:41 +00:00
@torch.no_grad()
def progressive_denoising(
self,
cond,
shape,
verbose=True,
callback=None,
quantize_denoised=False,
img_callback=None,
mask=None,
x0=None,
temperature=1.0,
noise_dropout=0.0,
score_corrector=None,
corrector_kwargs=None,
batch_size=None,
x_T=None,
start_T=None,
log_every_t=None,
):
2021-12-21 02:23:41 +00:00
if not log_every_t:
log_every_t = self.log_every_t
timesteps = self.num_timesteps
if batch_size is not None:
b = batch_size if batch_size is not None else shape[0]
shape = [batch_size] + list(shape)
else:
b = batch_size = shape[0]
if x_T is None:
img = torch.randn(shape, device=self.device)
else:
img = x_T
intermediates = []
if cond is not None:
if isinstance(cond, dict):
cond = {
key: cond[key][:batch_size]
if not isinstance(cond[key], list)
else list(map(lambda x: x[:batch_size], cond[key]))
for key in cond
}
2021-12-21 02:23:41 +00:00
else:
cond = (
[c[:batch_size] for c in cond]
if isinstance(cond, list)
else cond[:batch_size]
)
2021-12-21 02:23:41 +00:00
if start_T is not None:
timesteps = min(timesteps, start_T)
iterator = (
tqdm(
reversed(range(0, timesteps)),
desc='Progressive Generation',
total=timesteps,
)
if verbose
else reversed(range(0, timesteps))
)
2021-12-21 02:23:41 +00:00
if type(temperature) == float:
temperature = [temperature] * timesteps
for i in iterator:
ts = torch.full((b,), i, device=self.device, dtype=torch.long)
if self.shorten_cond_schedule:
assert self.model.conditioning_key != 'hybrid'
tc = self.cond_ids[ts].to(cond.device)
cond = self.q_sample(
x_start=cond, t=tc, noise=torch.randn_like(cond)
)
img, x0_partial = self.p_sample(
img,
cond,
ts,
clip_denoised=self.clip_denoised,
quantize_denoised=quantize_denoised,
return_x0=True,
temperature=temperature[i],
noise_dropout=noise_dropout,
score_corrector=score_corrector,
corrector_kwargs=corrector_kwargs,
)
2021-12-21 02:23:41 +00:00
if mask is not None:
assert x0 is not None
img_orig = self.q_sample(x0, ts)
img = img_orig * mask + (1.0 - mask) * img
2021-12-21 02:23:41 +00:00
if i % log_every_t == 0 or i == timesteps - 1:
intermediates.append(x0_partial)
if callback:
callback(i)
if img_callback:
img_callback(img, i)
2021-12-21 02:23:41 +00:00
return img, intermediates
@torch.no_grad()
def p_sample_loop(
self,
cond,
shape,
return_intermediates=False,
x_T=None,
verbose=True,
callback=None,
timesteps=None,
quantize_denoised=False,
mask=None,
x0=None,
img_callback=None,
start_T=None,
log_every_t=None,
):
2021-12-21 02:23:41 +00:00
if not log_every_t:
log_every_t = self.log_every_t
device = self.betas.device
b = shape[0]
if x_T is None:
img = torch.randn(shape, device=device)
else:
img = x_T
intermediates = [img]
if timesteps is None:
timesteps = self.num_timesteps
if start_T is not None:
timesteps = min(timesteps, start_T)
iterator = (
tqdm(
reversed(range(0, timesteps)),
desc='Sampling t',
total=timesteps,
)
if verbose
else reversed(range(0, timesteps))
)
2021-12-21 02:23:41 +00:00
if mask is not None:
assert x0 is not None
assert (
x0.shape[2:3] == mask.shape[2:3]
) # spatial size has to match
2021-12-21 02:23:41 +00:00
for i in iterator:
ts = torch.full((b,), i, device=device, dtype=torch.long)
if self.shorten_cond_schedule:
assert self.model.conditioning_key != 'hybrid'
tc = self.cond_ids[ts].to(cond.device)
cond = self.q_sample(
x_start=cond, t=tc, noise=torch.randn_like(cond)
)
img = self.p_sample(
img,
cond,
ts,
clip_denoised=self.clip_denoised,
quantize_denoised=quantize_denoised,
)
2021-12-21 02:23:41 +00:00
if mask is not None:
img_orig = self.q_sample(x0, ts)
img = img_orig * mask + (1.0 - mask) * img
2021-12-21 02:23:41 +00:00
if i % log_every_t == 0 or i == timesteps - 1:
intermediates.append(img)
if callback:
callback(i)
if img_callback:
img_callback(img, i)
2021-12-21 02:23:41 +00:00
if return_intermediates:
return img, intermediates
return img
@torch.no_grad()
def sample(
self,
cond,
batch_size=16,
return_intermediates=False,
x_T=None,
verbose=True,
timesteps=None,
quantize_denoised=False,
mask=None,
x0=None,
shape=None,
**kwargs,
):
2021-12-21 02:23:41 +00:00
if shape is None:
shape = (
batch_size,
self.channels,
self.image_size,
self.image_size,
)
2021-12-21 02:23:41 +00:00
if cond is not None:
if isinstance(cond, dict):
cond = {
key: cond[key][:batch_size]
if not isinstance(cond[key], list)
else list(map(lambda x: x[:batch_size], cond[key]))
for key in cond
}
2021-12-21 02:23:41 +00:00
else:
cond = (
[c[:batch_size] for c in cond]
if isinstance(cond, list)
else cond[:batch_size]
)
return self.p_sample_loop(
cond,
shape,
return_intermediates=return_intermediates,
x_T=x_T,
verbose=verbose,
timesteps=timesteps,
quantize_denoised=quantize_denoised,
mask=mask,
x0=x0,
)
2021-12-21 02:23:41 +00:00
@torch.no_grad()
def sample_log(self, cond, batch_size, ddim, ddim_steps, **kwargs):
if ddim:
ddim_sampler = DDIMSampler(self)
shape = (self.channels, self.image_size, self.image_size)
samples, intermediates = ddim_sampler.sample(
ddim_steps, batch_size, shape, cond, verbose=False, **kwargs
)
else:
samples, intermediates = self.sample(
cond=cond,
batch_size=batch_size,
return_intermediates=True,
**kwargs,
)
return samples, intermediates
@torch.no_grad()
def log_images(
self,
batch,
N=8,
n_row=4,
sample=True,
ddim_steps=200,
ddim_eta=1.0,
return_keys=None,
quantize_denoised=True,
inpaint=False,
plot_denoise_rows=False,
plot_progressive_rows=False,
plot_diffusion_rows=False,
**kwargs,
):
use_ddim = ddim_steps is not None
2021-12-21 02:23:41 +00:00
log = dict()
z, c, x, xrec, xc = self.get_input(
batch,
self.first_stage_key,
return_first_stage_outputs=True,
force_c_encode=True,
return_original_cond=True,
bs=N,
)
2021-12-21 02:23:41 +00:00
N = min(x.shape[0], N)
n_row = min(x.shape[0], n_row)
log['inputs'] = x
log['reconstruction'] = xrec
2021-12-21 02:23:41 +00:00
if self.model.conditioning_key is not None:
if hasattr(self.cond_stage_model, 'decode'):
2021-12-21 02:23:41 +00:00
xc = self.cond_stage_model.decode(c)
log['conditioning'] = xc
elif self.cond_stage_key in ['caption']:
xc = log_txt_as_img((x.shape[2], x.shape[3]), batch['caption'])
log['conditioning'] = xc
2021-12-21 02:23:41 +00:00
elif self.cond_stage_key == 'class_label':
xc = log_txt_as_img(
(x.shape[2], x.shape[3]), batch['human_label']
)
2021-12-21 02:23:41 +00:00
log['conditioning'] = xc
elif isimage(xc):
log['conditioning'] = xc
2021-12-21 02:23:41 +00:00
if ismap(xc):
log['original_conditioning'] = self.to_rgb(xc)
2021-12-21 02:23:41 +00:00
if plot_diffusion_rows:
# get diffusion row
diffusion_row = list()
z_start = z[:n_row]
for t in range(self.num_timesteps):
if t % self.log_every_t == 0 or t == self.num_timesteps - 1:
t = repeat(torch.tensor([t]), '1 -> b', b=n_row)
t = t.to(self.device).long()
noise = torch.randn_like(z_start)
z_noisy = self.q_sample(x_start=z_start, t=t, noise=noise)
diffusion_row.append(self.decode_first_stage(z_noisy))
diffusion_row = torch.stack(
diffusion_row
) # n_log_step, n_row, C, H, W
2021-12-21 02:23:41 +00:00
diffusion_grid = rearrange(diffusion_row, 'n b c h w -> b n c h w')
diffusion_grid = rearrange(
diffusion_grid, 'b n c h w -> (b n) c h w'
)
diffusion_grid = make_grid(
diffusion_grid, nrow=diffusion_row.shape[0]
)
log['diffusion_row'] = diffusion_grid
2021-12-21 02:23:41 +00:00
if sample:
# get denoise row
with self.ema_scope('Plotting'):
samples, z_denoise_row = self.sample_log(
cond=c,
batch_size=N,
ddim=use_ddim,
ddim_steps=ddim_steps,
eta=ddim_eta,
)
# samples, z_denoise_row = self.sample(cond=c, batch_size=N, return_intermediates=True)
2021-12-21 02:23:41 +00:00
x_samples = self.decode_first_stage(samples)
log['samples'] = x_samples
2021-12-21 02:23:41 +00:00
if plot_denoise_rows:
denoise_grid = self._get_denoise_row_from_list(z_denoise_row)
log['denoise_row'] = denoise_grid
uc = self.get_learned_conditioning(len(c) * [''])
sample_scaled, _ = self.sample_log(
cond=c,
batch_size=N,
ddim=use_ddim,
ddim_steps=ddim_steps,
eta=ddim_eta,
unconditional_guidance_scale=5.0,
unconditional_conditioning=uc,
)
log['samples_scaled'] = self.decode_first_stage(sample_scaled)
if (
quantize_denoised
and not isinstance(self.first_stage_model, AutoencoderKL)
and not isinstance(self.first_stage_model, IdentityFirstStage)
):
2021-12-21 02:23:41 +00:00
# also display when quantizing x0 while sampling
with self.ema_scope('Plotting Quantized Denoised'):
samples, z_denoise_row = self.sample_log(
cond=c,
batch_size=N,
ddim=use_ddim,
ddim_steps=ddim_steps,
eta=ddim_eta,
quantize_denoised=True,
)
# samples, z_denoise_row = self.sample(cond=c, batch_size=N, return_intermediates=True,
# quantize_denoised=True)
2021-12-21 02:23:41 +00:00
x_samples = self.decode_first_stage(samples.to(self.device))
log['samples_x0_quantized'] = x_samples
2021-12-21 02:23:41 +00:00
if inpaint:
# make a simple center square
b, h, w = z.shape[0], z.shape[2], z.shape[3]
mask = torch.ones(N, h, w).to(self.device)
# zeros will be filled in
mask[:, h // 4 : 3 * h // 4, w // 4 : 3 * w // 4] = 0.0
2021-12-21 02:23:41 +00:00
mask = mask[:, None, ...]
with self.ema_scope('Plotting Inpaint'):
samples, _ = self.sample_log(
cond=c,
batch_size=N,
ddim=use_ddim,
eta=ddim_eta,
ddim_steps=ddim_steps,
x0=z[:N],
mask=mask,
)
2021-12-21 02:23:41 +00:00
x_samples = self.decode_first_stage(samples.to(self.device))
log['samples_inpainting'] = x_samples
log['mask'] = mask
2021-12-21 02:23:41 +00:00
# outpaint
with self.ema_scope('Plotting Outpaint'):
samples, _ = self.sample_log(
cond=c,
batch_size=N,
ddim=use_ddim,
eta=ddim_eta,
ddim_steps=ddim_steps,
x0=z[:N],
mask=mask,
)
2021-12-21 02:23:41 +00:00
x_samples = self.decode_first_stage(samples.to(self.device))
log['samples_outpainting'] = x_samples
2021-12-21 02:23:41 +00:00
if plot_progressive_rows:
with self.ema_scope('Plotting Progressives'):
img, progressives = self.progressive_denoising(
c,
shape=(self.channels, self.image_size, self.image_size),
batch_size=N,
)
prog_row = self._get_denoise_row_from_list(
progressives, desc='Progressive Generation'
)
log['progressive_row'] = prog_row
2021-12-21 02:23:41 +00:00
if return_keys:
if np.intersect1d(list(log.keys()), return_keys).shape[0] == 0:
return log
else:
return {key: log[key] for key in return_keys}
return log
def configure_optimizers(self):
lr = self.learning_rate
2022-08-23 22:26:28 +00:00
if self.embedding_manager is not None:
params = list(self.embedding_manager.embedding_parameters())
# params = list(self.cond_stage_model.transformer.text_model.embeddings.embedding_manager.embedding_parameters())
else:
params = list(self.model.parameters())
if self.cond_stage_trainable:
print(
f'{self.__class__.__name__}: Also optimizing conditioner params!'
)
2022-08-23 22:26:28 +00:00
params = params + list(self.cond_stage_model.parameters())
if self.learn_logvar:
print('Diffusion model optimizing logvar')
params.append(self.logvar)
2021-12-21 02:23:41 +00:00
opt = torch.optim.AdamW(params, lr=lr)
if self.use_scheduler:
assert 'target' in self.scheduler_config
scheduler = instantiate_from_config(self.scheduler_config)
print('Setting up LambdaLR scheduler...')
2021-12-21 02:23:41 +00:00
scheduler = [
{
'scheduler': LambdaLR(opt, lr_lambda=scheduler.schedule),
'interval': 'step',
'frequency': 1,
}
]
2021-12-21 02:23:41 +00:00
return [opt], scheduler
return opt
@torch.no_grad()
def to_rgb(self, x):
x = x.float()
if not hasattr(self, 'colorize'):
2021-12-21 02:23:41 +00:00
self.colorize = torch.randn(3, x.shape[1], 1, 1).to(x)
x = nn.functional.conv2d(x, weight=self.colorize)
x = 2.0 * (x - x.min()) / (x.max() - x.min()) - 1.0
2021-12-21 02:23:41 +00:00
return x
2022-08-23 22:26:28 +00:00
@rank_zero_only
def on_save_checkpoint(self, checkpoint):
checkpoint.clear()
2022-08-23 22:26:28 +00:00
if os.path.isdir(self.trainer.checkpoint_callback.dirpath):
self.embedding_manager.save(
os.path.join(
self.trainer.checkpoint_callback.dirpath, 'embeddings.pt'
)
)
2022-08-23 22:26:28 +00:00
if (self.global_step - self.emb_ckpt_counter) > 500:
self.embedding_manager.save(
os.path.join(
self.trainer.checkpoint_callback.dirpath,
f'embeddings_gs-{self.global_step}.pt',
)
)
2022-08-23 22:26:28 +00:00
self.emb_ckpt_counter += 500
2021-12-21 02:23:41 +00:00
class DiffusionWrapper(pl.LightningModule):
def __init__(self, diff_model_config, conditioning_key):
super().__init__()
self.diffusion_model = instantiate_from_config(diff_model_config)
self.conditioning_key = conditioning_key
assert self.conditioning_key in [
None,
'concat',
'crossattn',
'hybrid',
'adm',
]
2021-12-21 02:23:41 +00:00
def forward(self, x, t, c_concat: list = None, c_crossattn: list = None):
if self.conditioning_key is None:
out = self.diffusion_model(x, t)
elif self.conditioning_key == 'concat':
xc = torch.cat([x] + c_concat, dim=1)
out = self.diffusion_model(xc, t)
elif self.conditioning_key == 'crossattn':
cc = torch.cat(c_crossattn, 1)
out = self.diffusion_model(x, t, context=cc)
elif self.conditioning_key == 'hybrid':
xc = torch.cat([x] + c_concat, dim=1)
cc = torch.cat(c_crossattn, 1)
out = self.diffusion_model(xc, t, context=cc)
elif self.conditioning_key == 'adm':
cc = c_crossattn[0]
out = self.diffusion_model(x, t, y=cc)
else:
raise NotImplementedError()
return out
class Layout2ImgDiffusion(LatentDiffusion):
# TODO: move all layout-specific hacks to this class
def __init__(self, cond_stage_key, *args, **kwargs):
assert (
cond_stage_key == 'coordinates_bbox'
), 'Layout2ImgDiffusion only for cond_stage_key="coordinates_bbox"'
2021-12-21 02:23:41 +00:00
super().__init__(cond_stage_key=cond_stage_key, *args, **kwargs)
def log_images(self, batch, N=8, *args, **kwargs):
logs = super().log_images(batch=batch, N=N, *args, **kwargs)
key = 'train' if self.training else 'validation'
dset = self.trainer.datamodule.datasets[key]
mapper = dset.conditional_builders[self.cond_stage_key]
bbox_imgs = []
map_fn = lambda catno: dset.get_textual_label(
dset.get_category_id(catno)
)
2021-12-21 02:23:41 +00:00
for tknzd_bbox in batch[self.cond_stage_key][:N]:
bboximg = mapper.plot(
tknzd_bbox.detach().cpu(), map_fn, (256, 256)
)
2021-12-21 02:23:41 +00:00
bbox_imgs.append(bboximg)
cond_img = torch.stack(bbox_imgs, dim=0)
logs['bbox_image'] = cond_img
return logs