Spaces:
Running
on
Zero
Running
on
Zero
import os | |
from contextlib import contextmanager, nullcontext | |
import torch | |
import wandb | |
from pytorch_lightning import LightningModule | |
from torch.nn.functional import mse_loss | |
from torch.nn.functional import sigmoid | |
from torch.optim import AdamW | |
from torch_ema import ExponentialMovingAverage as EMA | |
from torchmetrics.image.fid import FrechetInceptionDistance | |
from torchmetrics.image.inception import InceptionScore | |
from torchvision.transforms.functional import to_pil_image | |
from torchvision.utils import save_image | |
from utils.create_arch import create_arch | |
from huggingface_hub import PyTorchModelHubMixin | |
class MMSERectifiedFlow(LightningModule, | |
PyTorchModelHubMixin, | |
pipeline_tag="image-to-image", | |
license="mit", | |
): | |
def __init__(self, | |
stage, | |
arch, | |
conditional=False, | |
mmse_model_ckpt_path=None, | |
mmse_model_arch=None, | |
lr=5e-4, | |
weight_decay=1e-3, | |
betas=(0.9, 0.95), | |
mmse_noise_std=0.1, | |
num_flow_steps=50, | |
ema_decay=0.9999, | |
eps=0.0, | |
t_schedule='stratified_uniform', | |
*args, | |
**kwargs | |
): | |
super().__init__() | |
self.save_hyperparameters(logger=False) | |
if stage == 'flow': | |
if conditional: | |
condition_channels = 3 | |
else: | |
condition_channels = 0 | |
if mmse_model_arch is None and 'colorization' in kwargs and kwargs['colorization']: | |
condition_channels //= 3 | |
self.model = create_arch(arch, condition_channels) | |
self.mmse_model = create_arch(mmse_model_arch, 0) if mmse_model_arch is not None else None | |
if mmse_model_ckpt_path is not None: | |
ckpt = torch.load(mmse_model_ckpt_path, map_location="cpu") | |
if mmse_model_arch is None: | |
mmse_model_arch = ckpt['hyper_parameters']['arch'] | |
self.mmse_model = create_arch(mmse_model_arch, 0) | |
if 'ema' in ckpt: | |
# ema_decay doesn't affect anything here, because we are doing load_state_dict | |
mmse_ema = EMA(self.mmse_model.parameters(), decay=ema_decay) | |
mmse_ema.load_state_dict(ckpt['ema']) | |
mmse_ema.copy_to() | |
elif 'params_ema' in ckpt: | |
self.mmse_model.load_state_dict(ckpt['params_ema']) | |
else: | |
state_dict = ckpt['state_dict'] | |
state_dict = {layer_name.replace('model.', ''): weights for layer_name, weights in | |
state_dict.items()} | |
state_dict = {layer_name.replace('module.', ''): weights for layer_name, weights in | |
state_dict.items()} | |
self.mmse_model.load_state_dict(state_dict) | |
for param in self.mmse_model.parameters(): | |
param.requires_grad = False | |
self.mmse_model.eval() | |
else: | |
assert stage == 'mmse' or stage == 'naive_flow' | |
assert not conditional | |
self.model = create_arch(arch, 0) | |
self.mmse_model = None | |
if 'flow' in stage: | |
self.fid = FrechetInceptionDistance(reset_real_features=True, normalize=True) | |
self.inception_score = InceptionScore(normalize=True) | |
self.ema = EMA(self.model.parameters(), decay=ema_decay) if self.ema_wanted else None | |
self.test_results_path = None | |
def ema_wanted(self): | |
return self.hparams.ema_decay != -1 | |
def on_save_checkpoint(self, checkpoint: dict) -> None: | |
if self.ema_wanted: | |
checkpoint['ema'] = self.ema.state_dict() | |
return super().on_save_checkpoint(checkpoint) | |
def on_load_checkpoint(self, checkpoint: dict) -> None: | |
if self.ema_wanted: | |
self.ema.load_state_dict(checkpoint['ema']) | |
return super().on_load_checkpoint(checkpoint) | |
def on_before_zero_grad(self, optimizer) -> None: | |
if self.ema_wanted: | |
self.ema.update(self.model.parameters()) | |
return super().on_before_zero_grad(optimizer) | |
def to(self, *args, **kwargs): | |
if self.ema_wanted: | |
self.ema.to(*args, **kwargs) | |
return super().to(*args, **kwargs) | |
# This will use the contextmanager of ema, to copy the EMA weights to the flow model during validation, and then restore them for training. | |
def maybe_ema(self): | |
ema = self.ema | |
ctx = nullcontext if ema is None else ema.average_parameters | |
yield ctx | |
def forward_mmse(self, y): | |
return self.model(y).clip(0, 1) | |
def forward_flow(self, x_t, t, y=None): | |
if self.hparams.conditional: | |
if self.mmse_model is not None: | |
with torch.no_grad(): | |
self.mmse_model.eval() | |
condition = self.mmse_model(y).clip(0, 1) | |
else: | |
condition = y | |
x_t = torch.cat((x_t, condition), dim=1) | |
return self.model(x_t, t) | |
def forward(self, x_t, t, y): | |
if 'flow' in self.hparams.stage: | |
return self.forward_flow(x_t, t, y) | |
else: | |
return self.forward_mmse(y) | |
def create_source_distribution_samples(self, x, y, non_noisy_z0): | |
with torch.no_grad(): | |
if self.hparams.conditional: | |
source_dist_samples = torch.randn_like(x) | |
else: | |
if self.hparams.stage == 'flow': | |
if non_noisy_z0 is None: | |
self.mmse_model.eval() | |
non_noisy_z0 = self.mmse_model(y).clip(0, 1) | |
source_dist_samples = non_noisy_z0 + torch.randn_like(non_noisy_z0) * self.hparams.mmse_noise_std | |
else: | |
assert self.hparams.stage == 'naive_flow' | |
if non_noisy_z0 is not None: | |
source_dist_samples = non_noisy_z0 | |
else: | |
source_dist_samples = y | |
if source_dist_samples.shape[1] != x.shape[1]: | |
assert source_dist_samples.shape[1] == 1 # Colorization | |
source_dist_samples = source_dist_samples.expand(-1, x.shape[1], -1, -1) | |
if self.hparams.mmse_noise_std is not None: | |
source_dist_samples = source_dist_samples + torch.randn_like(source_dist_samples) * self.hparams.mmse_noise_std | |
return source_dist_samples | |
def stratified_uniform(bs, group=0, groups=1, dtype=None, device=None): | |
if groups <= 0: | |
raise ValueError(f"groups must be positive, got {groups}") | |
if group < 0 or group >= groups: | |
raise ValueError(f"group must be in [0, {groups})") | |
n = bs * groups | |
offsets = torch.arange(group, n, groups, dtype=dtype, device=device) | |
u = torch.rand(bs, dtype=dtype, device=device) | |
return ((offsets + u) / n).view(bs, 1, 1, 1) | |
def generate_random_t(self, bs, dtype=None): | |
if self.hparams.t_schedule == 'logit-normal': | |
return sigmoid(torch.randn(bs, 1, 1, 1, device=self.device)) * (1.0 - self.hparams.eps) + self.hparams.eps | |
elif self.hparams.t_schedule == 'uniform': | |
return torch.rand(bs, 1, 1, 1, device=self.device) * (1.0 - self.hparams.eps) + self.hparams.eps | |
elif self.hparams.t_schedule == 'stratified_uniform': | |
return self.stratified_uniform(bs, self.trainer.global_rank, self.trainer.world_size, dtype=dtype, | |
device=self.device) * (1.0 - self.hparams.eps) + self.hparams.eps | |
else: | |
raise NotImplementedError() | |
def training_step(self, batch, batch_idx): | |
x = batch['x'] | |
y = batch['y'] | |
non_noisy_z0 = batch['non_noisy_z0'] if 'non_noisy_z0' in batch else None | |
if 'flow' in self.hparams.stage: | |
with torch.no_grad(): | |
t = self.generate_random_t(x.shape[0], dtype=x.dtype) | |
source_dist_samples = self.create_source_distribution_samples(x, y, non_noisy_z0) | |
x_t = t * x + (1.0 - t) * source_dist_samples | |
v_t = self(x_t, t.squeeze(), y) | |
loss = mse_loss(v_t, x - source_dist_samples) | |
else: | |
xhat = self(x_t=None, t=None, y=y) | |
loss = mse_loss(xhat, x) | |
self.log("train/loss", loss) | |
return loss | |
def generate_reconstructions(self, x, y, non_noisy_z0, num_flow_steps, result_device): | |
with self.maybe_ema(): | |
if 'flow' in self.hparams.stage: | |
source_dist_samples = self.create_source_distribution_samples(x, y, non_noisy_z0) | |
dt = (1.0 / num_flow_steps) * (1.0 - self.hparams.eps) | |
x_t_next = source_dist_samples.clone() | |
x_t_seq = [x_t_next] | |
t_one = torch.ones(x.shape[0], device=self.device) | |
for i in range(num_flow_steps): | |
num_t = (i / num_flow_steps) * (1.0 - self.hparams.eps) + self.hparams.eps | |
v_t_next = self(x_t=x_t_next, t=t_one * num_t, y=y).to(x_t_next.dtype) | |
x_t_next = x_t_next.clone() + v_t_next * dt | |
x_t_seq.append(x_t_next.to(result_device)) | |
xhat = x_t_seq[-1].clip(0, 1).to(torch.float32) | |
source_dist_samples = source_dist_samples.to(result_device) | |
else: | |
xhat = self(x_t=None, t=None, y=y).to(torch.float32) | |
x_t_seq = None | |
source_dist_samples = None | |
return xhat.to(result_device), x_t_seq, source_dist_samples | |
def validation_step(self, batch, batch_idx): | |
x = batch['x'] | |
y = batch['y'] | |
non_noisy_z0 = batch['non_noisy_z0'] if 'non_noisy_z0' in batch else None | |
xhat, x_t_seq, source_dist_samples = self.generate_reconstructions(x, y, non_noisy_z0, self.hparams.num_flow_steps, | |
self.device) | |
x = x.to(torch.float32) | |
y = y.to(torch.float32) | |
self.log_dict({"val_metrics/mse": ((x - xhat) ** 2).mean()}, on_step=False, on_epoch=True, sync_dist=True, | |
batch_size=x.shape[0]) | |
if 'flow' in self.hparams.stage: | |
self.fid.update(x, real=True) | |
self.fid.update(xhat, real=False) | |
self.inception_score.update(xhat) | |
if batch_idx == 0: | |
wandb_logger = self.logger.experiment | |
wandb_logger.log({'val_images/x': [wandb.Image(to_pil_image(create_grid(x)))], | |
'val_images/y': [wandb.Image(to_pil_image(create_grid(y.clip(0, 1))))], | |
'val_images/xhat': [wandb.Image(to_pil_image(create_grid(xhat)))], }) | |
if 'flow' in self.hparams.stage: | |
wandb_logger.log({'val_images/x_t_seq': [wandb.Image(to_pil_image(create_grid( | |
torch.cat([elem[0].unsqueeze(0).to(torch.float32) for elem in x_t_seq], dim=0).clip(0, 1), | |
num_images=len(x_t_seq))))], 'val_images/source_distribution_samples': [ | |
wandb.Image(to_pil_image(create_grid(source_dist_samples.clip(0, 1).to(torch.float32))))]}) | |
if self.mmse_model is not None: | |
xhat_mmse = self.mmse_model(y).clip(0, 1) | |
wandb_logger.log({'val_images/xhat_mmse': [ | |
wandb.Image(to_pil_image(create_grid(xhat_mmse.to(torch.float32))))]}) | |
def on_validation_epoch_end(self): | |
if 'flow' in self.hparams.stage: | |
inception_score_mean, inception_score_std = self.inception_score.compute() | |
self.log_dict( | |
{'val_metrics/fid': self.fid.compute(), | |
'val_metrics/inception_score_mean': inception_score_mean, | |
'val_metrics/inception_score_std': inception_score_std}, | |
on_epoch=True, on_step=False, sync_dist=True, | |
batch_size=1) | |
self.fid.reset() | |
self.inception_score.reset() | |
def test_step(self, batch, batch_idx): | |
assert self.test_results_path is not None, "Please set test_results_path before testing." | |
assert os.path.isdir(self.test_results_path), 'Please make sure the test_result_path dir exists.' | |
def save_image_batch(images, folder, image_file_names): | |
os.makedirs(folder, exist_ok=True) | |
for i, img in enumerate(images): | |
save_image(images[i].clip(0, 1), os.path.join(folder, image_file_names[i])) | |
os.makedirs(self.test_results_path, exist_ok=True) | |
x = batch['x'] | |
y = batch['y'] | |
non_noisy_z0 = batch['non_noisy_z0'] if 'non_noisy_z0' in batch else None | |
y_path = os.path.join(self.test_results_path, 'y') | |
save_image_batch(y, y_path, batch['img_file_name']) | |
if 'flow' in self.hparams.stage: | |
source_dist_samples_to_save = None | |
for num_flow_steps in self.num_test_flow_steps: | |
xhat, x_t_seq, source_dist_samples = self.generate_reconstructions(x, y, non_noisy_z0, num_flow_steps, | |
torch.device("cpu")) | |
xhat_path = os.path.join(self.test_results_path, f"num_flow_steps={num_flow_steps}", 'xhat') | |
save_image_batch(xhat, xhat_path, batch['img_file_name']) | |
if source_dist_samples_to_save is None: | |
source_dist_samples_to_save = source_dist_samples | |
source_distribution_samples_path = os.path.join(self.test_results_path, 'source_distribution_samples') | |
save_image_batch(source_dist_samples_to_save, source_distribution_samples_path, batch['img_file_name']) | |
if self.mmse_model is not None: | |
mmse_estimates = self.mmse_model(y).clip(0, 1) | |
mmse_samples_path = os.path.join(self.test_results_path, 'mmse_samples') | |
save_image_batch(mmse_estimates, mmse_samples_path, batch['img_file_name']) | |
else: | |
xhat, _, _ = self.generate_reconstructions(x, y, non_noisy_z0, None, torch.device('cpu')) | |
xhat_path = os.path.join(self.test_results_path, 'xhat') | |
save_image_batch(xhat, xhat_path, batch['img_file_name']) | |
def configure_optimizers(self): | |
# Add here a learning rate scheduler if you wish to do so. | |
optimizer = AdamW(self.model.parameters(), | |
betas=self.hparams.betas, | |
eps=1e-8, | |
lr=self.hparams.lr, | |
weight_decay=self.hparams.weight_decay) | |
return optimizer | |