""" wild mixture of https://github.com/openai/improved-diffusion/blob/e94489283bb876ac1477d5dd7709bbbd2d9902ce/improved_diffusion/gaussian_diffusion.py https://github.com/lucidrains/denoising-diffusion-pytorch/blob/7706bdfc6f527f58d33f84b7b522e61e6e3164b3/denoising_diffusion_pytorch/denoising_diffusion_pytorch.py https://github.com/CompVis/taming-transformers -- merci """ from functools import partial from contextlib import contextmanager import numpy as np from tqdm import tqdm from einops import rearrange, repeat import logging mainlogger = logging.getLogger('mainlogger') import random import torch import torch.nn as nn from torch.optim.lr_scheduler import LambdaLR, CosineAnnealingLR from torchvision.utils import make_grid import pytorch_lightning as pl from pytorch_lightning.utilities import rank_zero_only from utils.utils import instantiate_from_config from lvdm.ema import LitEma from lvdm.models.samplers.ddim import DDIMSampler from lvdm.distributions import DiagonalGaussianDistribution from lvdm.models.utils_diffusion import make_beta_schedule, rescale_zero_terminal_snr from lvdm.basics import disabled_train from lvdm.common import ( extract_into_tensor, noise_like, exists, default ) import math from lvdm.models.autoencoder_dualref import VideoDecoder __conditioning_keys__ = {'concat': 'c_concat', 'crossattn': 'c_crossattn', 'adm': 'y'} 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=None, 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., v_posterior=0., # weight for choosing posterior variance as sigma = (1-v) * beta_tilde + v * beta l_simple_weight=1., conditioning_key=None, parameterization="eps", # all assuming fixed variance schedules scheduler_config=None, use_positional_encodings=False, learn_logvar=False, logvar_init=0., rescale_betas_zero_snr=False, ): super().__init__() assert parameterization in ["eps", "x0", "v"], 'currently only supporting "eps" and "x0" and "v"' self.parameterization = parameterization mainlogger.info(f"{self.__class__.__name__}: Running in {self.parameterization}-prediction mode") 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.channels = channels self.temporal_length = unet_config.params.temporal_length self.image_size = image_size # try conv? if isinstance(self.image_size, int): self.image_size = [self.image_size, self.image_size] 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 self.rescale_betas_zero_snr = rescale_betas_zero_snr if self.use_ema: self.model_ema = LitEma(self.model) mainlogger.info(f"Keeping EMAs of {len(list(self.model_ema.buffers()))}.") 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 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) ## for reschedule self.given_betas = given_betas self.beta_schedule = beta_schedule self.timesteps = timesteps self.cosine_s = cosine_s self.loss_type = loss_type self.learn_logvar = learn_logvar self.logvar = torch.full(fill_value=logvar_init, size=(self.num_timesteps,)) 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): 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) if self.rescale_betas_zero_snr: betas = rescale_zero_terminal_snr(betas) alphas = 1. - betas alphas_cumprod = np.cumprod(alphas, axis=0) alphas_cumprod_prev = np.append(1., alphas_cumprod[:-1]) timesteps, = betas.shape 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' 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)) # 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. - alphas_cumprod))) self.register_buffer('log_one_minus_alphas_cumprod', to_torch(np.log(1. - alphas_cumprod))) if self.parameterization != 'v': self.register_buffer('sqrt_recip_alphas_cumprod', to_torch(np.sqrt(1. / alphas_cumprod))) self.register_buffer('sqrt_recipm1_alphas_cumprod', to_torch(np.sqrt(1. / alphas_cumprod - 1))) else: self.register_buffer('sqrt_recip_alphas_cumprod', torch.zeros_like(to_torch(alphas_cumprod))) self.register_buffer('sqrt_recipm1_alphas_cumprod', torch.zeros_like(to_torch(alphas_cumprod))) # calculations for posterior q(x_{t-1} | x_t, x_0) posterior_variance = (1 - self.v_posterior) * betas * (1. - alphas_cumprod_prev) / ( 1. - alphas_cumprod) + self.v_posterior * betas # above: equal to 1. / (1. / (1. - alpha_cumprod_tm1) + alpha_t / beta_t) self.register_buffer('posterior_variance', to_torch(posterior_variance)) # 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. - alphas_cumprod))) self.register_buffer('posterior_mean_coef2', to_torch( (1. - alphas_cumprod_prev) * np.sqrt(alphas) / (1. - 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. * 1 - torch.Tensor(alphas_cumprod)) elif self.parameterization == "v": lvlb_weights = torch.ones_like(self.betas ** 2 / ( 2 * self.posterior_variance * to_torch(alphas) * (1 - self.alphas_cumprod))) else: raise NotImplementedError("mu not supported") # 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: mainlogger.info(f"{context}: Switched to EMA weights") try: yield None finally: if self.use_ema: self.model_ema.restore(self.model.parameters()) if context is not None: mainlogger.info(f"{context}: Restored training weights") 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"] keys = list(sd.keys()) for k in keys: for ik in ignore_keys: if k.startswith(ik): mainlogger.info("Deleting key {} from state_dict.".format(k)) 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) mainlogger.info(f"Restored from {path} with {len(missing)} missing and {len(unexpected)} unexpected keys") if len(missing) > 0: mainlogger.info(f"Missing Keys: {missing}") if len(unexpected) > 0: mainlogger.info(f"Unexpected Keys: {unexpected}") 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) 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 ) def predict_start_from_z_and_v(self, x_t, t, v): # 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. - alphas_cumprod))) return ( extract_into_tensor(self.sqrt_alphas_cumprod, t, x_t.shape) * x_t - extract_into_tensor(self.sqrt_one_minus_alphas_cumprod, t, x_t.shape) * v ) def predict_eps_from_z_and_v(self, x_t, t, v): return ( extract_into_tensor(self.sqrt_alphas_cumprod, t, x_t.shape) * v + extract_into_tensor(self.sqrt_one_minus_alphas_cumprod, t, x_t.shape) * x_t ) 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 def p_mean_variance(self, x, t, clip_denoised: bool): model_out = self.model(x, t) if self.parameterization == "eps": x_recon = self.predict_start_from_noise(x, t=t, noise=model_out) elif self.parameterization == "x0": x_recon = model_out if clip_denoised: x_recon.clamp_(-1., 1.) model_mean, posterior_variance, posterior_log_variance = self.q_posterior(x_start=x_recon, x_t=x, t=t) 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) 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 @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): img = self.p_sample(img, torch.full((b,), i, device=device, dtype=torch.long), clip_denoised=self.clip_denoised) 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) 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) def get_v(self, x, noise, t): return ( extract_into_tensor(self.sqrt_alphas_cumprod, t, x.shape) * noise - extract_into_tensor(self.sqrt_one_minus_alphas_cumprod, t, x.shape) * x ) 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') 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": target = noise elif self.parameterization == "x0": target = x_start elif self.parameterization == "v": target = self.get_v(x_start, noise, t) else: raise NotImplementedError(f"Paramterization {self.parameterization} not yet supported") 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() 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) self.log("global_step", self.global_step, prog_bar=True, logger=True, on_step=True, on_epoch=False) 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) 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) 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): 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 # 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) if sample: # get denoise row with self.ema_scope("Plotting"): samples, denoise_row = self.sample(batch_size=N, return_intermediates=True) log["samples"] = samples log["denoise_row"] = self._get_rows_from_list(denoise_row) 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, num_timesteps_cond=None, cond_stage_key="caption", cond_stage_trainable=False, cond_stage_forward=None, conditioning_key=None, uncond_prob=0.2, uncond_type="empty_seq", scale_factor=1.0, scale_by_std=False, encoder_type="2d", only_model=False, noise_strength=0, use_dynamic_rescale=False, base_scale=0.7, turning_step=400, loop_video=False, fps_condition_type='fs', perframe_ae=False, # added logdir=None, rand_cond_frame=False, en_and_decode_n_samples_a_time=None, *args, **kwargs): 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 ckpt_path = kwargs.pop("ckpt_path", None) ignore_keys = kwargs.pop("ignore_keys", []) conditioning_key = default(conditioning_key, 'crossattn') super().__init__(conditioning_key=conditioning_key, *args, **kwargs) self.cond_stage_trainable = cond_stage_trainable self.cond_stage_key = cond_stage_key self.noise_strength = noise_strength self.use_dynamic_rescale = use_dynamic_rescale self.loop_video = loop_video self.fps_condition_type = fps_condition_type self.perframe_ae = perframe_ae self.logdir = logdir self.rand_cond_frame = rand_cond_frame self.en_and_decode_n_samples_a_time = en_and_decode_n_samples_a_time try: self.num_downs = len(first_stage_config.params.ddconfig.ch_mult) - 1 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)) if use_dynamic_rescale: scale_arr1 = np.linspace(1.0, base_scale, turning_step) scale_arr2 = np.full(self.num_timesteps, base_scale) scale_arr = np.concatenate((scale_arr1, scale_arr2)) to_torch = partial(torch.tensor, dtype=torch.float32) self.register_buffer('scale_arr', to_torch(scale_arr)) self.instantiate_first_stage(first_stage_config) self.instantiate_cond_stage(cond_stage_config) self.first_stage_config = first_stage_config self.cond_stage_config = cond_stage_config self.clip_denoised = False self.cond_stage_forward = cond_stage_forward self.encoder_type = encoder_type assert(encoder_type in ["2d", "3d"]) self.uncond_prob = uncond_prob self.classifier_free_guidance = True if uncond_prob > 0 else False assert(uncond_type in ["zero_embed", "empty_seq"]) self.uncond_type = uncond_type self.restarted_from_ckpt = False if ckpt_path is not None: self.init_from_ckpt(ckpt_path, ignore_keys, only_model=only_model) self.restarted_from_ckpt = 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 @rank_zero_only @torch.no_grad() def on_train_batch_start(self, batch, batch_idx, dataloader_idx=None): # only for very first batch, reset the self.scale_factor 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., 'rather not use custom rescaling and std-rescaling simultaneously' # set rescale weight to 1./std of encodings mainlogger.info("### USING STD-RESCALING ###") 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. / z.flatten().std()) mainlogger.info(f"setting self.scale_factor to {self.scale_factor}") mainlogger.info("### USING STD-RESCALING ###") mainlogger.info(f"std={z.flatten().std()}") 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) 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: 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: model = instantiate_from_config(config) self.cond_stage_model = model 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) 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 get_first_stage_encoding(self, encoder_posterior, noise=None): if isinstance(encoder_posterior, DiagonalGaussianDistribution): z = encoder_posterior.sample(noise=noise) elif isinstance(encoder_posterior, torch.Tensor): z = encoder_posterior else: raise NotImplementedError(f"encoder_posterior of type '{type(encoder_posterior)}' not yet implemented") return self.scale_factor * z @torch.no_grad() def encode_first_stage(self, x): if self.encoder_type == "2d" and x.dim() == 5: b, _, t, _, _ = x.shape x = rearrange(x, 'b c t h w -> (b t) c h w') reshape_back = True else: reshape_back = False ## consume more GPU memory but faster if not self.perframe_ae: encoder_posterior = self.first_stage_model.encode(x) results = self.get_first_stage_encoding(encoder_posterior).detach() else: ## consume less GPU memory but slower results = [] for index in range(x.shape[0]): frame_batch = self.first_stage_model.encode(x[index:index+1,:,:,:]) frame_result = self.get_first_stage_encoding(frame_batch).detach() results.append(frame_result) results = torch.cat(results, dim=0) if reshape_back: results = rearrange(results, '(b t) c h w -> b c t h w', b=b,t=t) return results def decode_core(self, z, **kwargs): if self.encoder_type == "2d" and z.dim() == 5: b, _, t, _, _ = z.shape z = rearrange(z, 'b c t h w -> (b t) c h w') reshape_back = True else: reshape_back = False z = 1. / self.scale_factor * z if not self.perframe_ae: results = self.first_stage_model.decode(z, **kwargs) else: results = [] n_samples = default(self.en_and_decode_n_samples_a_time, self.temporal_length) n_rounds = math.ceil(z.shape[0] / n_samples) with torch.autocast("cuda", enabled=True): for n in range(n_rounds): if isinstance(self.first_stage_model.decoder, VideoDecoder): kwargs.update({"timesteps": len(z[n * n_samples : (n + 1) * n_samples])}) else: kwargs = {} out = self.first_stage_model.decode( z[n * n_samples : (n + 1) * n_samples], **kwargs ) results.append(out) results = torch.cat(results, dim=0) if reshape_back: results = rearrange(results, '(b t) c h w -> b c t h w', b=b,t=t) return results @torch.no_grad() def decode_first_stage(self, z, **kwargs): return self.decode_core(z, **kwargs) # same as above but without decorator def differentiable_decode_first_stage(self, z, **kwargs): return self.decode_core(z, **kwargs) @torch.no_grad() def get_batch_input(self, batch, random_uncond, return_first_stage_outputs=False, return_original_cond=False): ## video shape: b, c, t, h, w x = super().get_input(batch, self.first_stage_key) ## encode video frames x to z via a 2D encoder z = self.encode_first_stage(x) ## get caption condition cond = batch[self.cond_stage_key] if random_uncond and self.uncond_type == 'empty_seq': for i, ci in enumerate(cond): if random.random() < self.uncond_prob: cond[i] = "" if isinstance(cond, dict) or isinstance(cond, list): cond_emb = self.get_learned_conditioning(cond) else: cond_emb = self.get_learned_conditioning(cond.to(self.device)) if random_uncond and self.uncond_type == 'zero_embed': for i, ci in enumerate(cond): if random.random() < self.uncond_prob: cond_emb[i] = torch.zeros_like(cond_emb[i]) out = [z, cond_emb] ## optional output: self-reconst or caption if return_first_stage_outputs: xrec = self.decode_first_stage(z) out.extend([xrec]) if return_original_cond: out.append(cond) return out def forward(self, x, c, **kwargs): t = torch.randint(0, self.num_timesteps, (x.shape[0],), device=self.device).long() if self.use_dynamic_rescale: x = x * extract_into_tensor(self.scale_arr, t, x.shape) return self.p_losses(x, c, t, **kwargs) def shared_step(self, batch, random_uncond, **kwargs): x, c = self.get_batch_input(batch, random_uncond=random_uncond) loss, loss_dict = self(x, c, **kwargs) return loss, loss_dict def apply_model(self, x_noisy, t, cond, **kwargs): 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' cond = {key: cond} x_recon = self.model(x_noisy, t, **cond, **kwargs) if isinstance(x_recon, tuple): return x_recon[0] else: return x_recon def p_losses(self, x_start, cond, t, noise=None, **kwargs): if self.noise_strength > 0: b, c, f, _, _ = x_start.shape offset_noise = torch.randn(b, c, f, 1, 1, device=x_start.device) noise = default(noise, lambda: torch.randn_like(x_start) + self.noise_strength * offset_noise) else: 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, **kwargs) loss_dict = {} prefix = 'train' if self.training else 'val' if self.parameterization == "x0": target = x_start elif self.parameterization == "eps": target = noise elif self.parameterization == "v": target = self.get_v(x_start, noise, t) else: raise NotImplementedError() loss_simple = self.get_loss(model_output, target, mean=False).mean([1, 2, 3, 4]) loss_dict.update({f'{prefix}/loss_simple': loss_simple.mean()}) if self.logvar.device is not self.device: self.logvar = self.logvar.to(self.device) logvar_t = self.logvar[t] # logvar_t = self.logvar[t.item()].to(self.device) # device conflict when ddp shared 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, 4)) 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) loss_dict.update({f'{prefix}/loss': loss}) return loss, loss_dict def training_step(self, batch, batch_idx): loss, loss_dict = self.shared_step(batch, random_uncond=self.classifier_free_guidance) ## sync_dist | rank_zero_only self.log_dict(loss_dict, prog_bar=True, logger=True, on_step=True, on_epoch=True, sync_dist=False) #self.log("epoch/global_step", self.global_step.float(), prog_bar=True, logger=True, on_step=True, on_epoch=False) ''' 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, rank_zero_only=True) ''' if (batch_idx+1) % self.log_every_t == 0: mainlogger.info(f"batch:{batch_idx}|epoch:{self.current_epoch} [globalstep:{self.global_step}]: loss={loss}") return loss def _get_denoise_row_from_list(self, samples, desc=''): denoise_row = [] for zd in tqdm(samples, desc=desc): denoise_row.append(self.decode_first_stage(zd.to(self.device))) n_log_timesteps = len(denoise_row) denoise_row = torch.stack(denoise_row) # n_log_timesteps, b, C, H, W if denoise_row.dim() == 5: 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_log_timesteps) elif denoise_row.dim() == 6: # video, grid_size=[n_log_timesteps*bs, t] video_length = denoise_row.shape[3] denoise_grid = rearrange(denoise_row, 'n b c t h w -> b n c t h w') denoise_grid = rearrange(denoise_grid, 'b n c t h w -> (b n) c t h w') denoise_grid = rearrange(denoise_grid, 'n c t h w -> (n t) c h w') denoise_grid = make_grid(denoise_grid, nrow=video_length) else: raise ValueError return denoise_grid @torch.no_grad() def log_images(self, batch, sample=True, ddim_steps=200, ddim_eta=1., plot_denoise_rows=False, \ unconditional_guidance_scale=1.0, **kwargs): """ log images for LatentDiffusion """ ##### control sampled imgae for logging, larger value may cause OOM sampled_img_num = 2 for key in batch.keys(): batch[key] = batch[key][:sampled_img_num] ## TBD: currently, classifier_free_guidance sampling is only supported by DDIM use_ddim = ddim_steps is not None log = dict() z, c, xrec, xc = self.get_batch_input(batch, random_uncond=False, return_first_stage_outputs=True, return_original_cond=True) N = xrec.shape[0] log["reconst"] = xrec log["condition"] = xc if sample: # get uncond embedding for classifier-free guidance sampling if unconditional_guidance_scale != 1.0: if isinstance(c, dict): c_cat, c_emb = c["c_concat"][0], c["c_crossattn"][0] log["condition_cat"] = c_cat else: c_emb = c if self.uncond_type == "empty_seq": prompts = N * [""] uc = self.get_learned_conditioning(prompts) elif self.uncond_type == "zero_embed": uc = torch.zeros_like(c_emb) ## hybrid case if isinstance(c, dict): uc_hybrid = {"c_concat": [c_cat], "c_crossattn": [uc]} uc = uc_hybrid else: uc = None 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, unconditional_guidance_scale=unconditional_guidance_scale, unconditional_conditioning=uc, x0=z, **kwargs) x_samples = self.decode_first_stage(samples) log["samples"] = x_samples if plot_denoise_rows: denoise_grid = self._get_denoise_row_from_list(z_denoise_row) log["denoise_row"] = denoise_grid return log def p_mean_variance(self, x, c, t, clip_denoised: bool, return_x0=False, score_corrector=None, corrector_kwargs=None, **kwargs): t_in = t model_out = self.apply_model(x, t_in, c, **kwargs) if score_corrector is not None: assert self.parameterization == "eps" model_out = score_corrector.modify_score(self, model_out, x, t, c, **corrector_kwargs) if self.parameterization == "eps": x_recon = self.predict_start_from_noise(x, t=t, noise=model_out) elif self.parameterization == "x0": x_recon = model_out else: raise NotImplementedError() if clip_denoised: x_recon.clamp_(-1., 1.) model_mean, posterior_variance, posterior_log_variance = self.q_posterior(x_start=x_recon, x_t=x, t=t) if return_x0: return model_mean, posterior_variance, posterior_log_variance, x_recon 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_x0=False, \ temperature=1., noise_dropout=0., score_corrector=None, corrector_kwargs=None, **kwargs): b, *_, device = *x.shape, x.device outputs = self.p_mean_variance(x=x, c=c, t=t, clip_denoised=clip_denoised, return_x0=return_x0, \ score_corrector=score_corrector, corrector_kwargs=corrector_kwargs, **kwargs) if 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.: 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))) if return_x0: return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise, x0 else: return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise @torch.no_grad() def p_sample_loop(self, cond, shape, return_intermediates=False, x_T=None, verbose=True, callback=None, \ timesteps=None, mask=None, x0=None, img_callback=None, start_T=None, log_every_t=None, **kwargs): if not log_every_t: log_every_t = self.log_every_t device = self.betas.device b = shape[0] # sample an initial noise 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)) if mask is not None: assert x0 is not None assert x0.shape[2:3] == mask.shape[2:3] # spatial size has to match 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, **kwargs) if mask is not None: img_orig = self.q_sample(x0, ts) img = img_orig * mask + (1. - mask) * img if i % log_every_t == 0 or i == timesteps - 1: intermediates.append(img) if callback: callback(i) if img_callback: img_callback(img, i) 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, mask=None, x0=None, shape=None, **kwargs): if shape is None: shape = (batch_size, self.channels, self.temporal_length, *self.image_size) 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} 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, mask=mask, x0=x0, **kwargs) @torch.no_grad() def sample_log(self, cond, batch_size, ddim, ddim_steps, **kwargs): if ddim: ddim_sampler = DDIMSampler(self) shape = (self.channels, self.temporal_length, *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 def configure_schedulers(self, optimizer): assert 'target' in self.scheduler_config scheduler_name = self.scheduler_config.target.split('.')[-1] interval = self.scheduler_config.interval frequency = self.scheduler_config.frequency if scheduler_name == "LambdaLRScheduler": scheduler = instantiate_from_config(self.scheduler_config) scheduler.start_step = self.global_step lr_scheduler = { 'scheduler': LambdaLR(optimizer, lr_lambda=scheduler.schedule), 'interval': interval, 'frequency': frequency } elif scheduler_name == "CosineAnnealingLRScheduler": scheduler = instantiate_from_config(self.scheduler_config) decay_steps = scheduler.decay_steps last_step = -1 if self.global_step == 0 else scheduler.start_step lr_scheduler = { 'scheduler': CosineAnnealingLR(optimizer, T_max=decay_steps, last_epoch=last_step), 'interval': interval, 'frequency': frequency } else: raise NotImplementedError return lr_scheduler class LatentVisualDiffusion(LatentDiffusion): def __init__(self, img_cond_stage_config, image_proj_stage_config, freeze_embedder=True, image_proj_model_trainable=True, *args, **kwargs): super().__init__(*args, **kwargs) self.image_proj_model_trainable = image_proj_model_trainable self._init_embedder(img_cond_stage_config, freeze_embedder) self._init_img_ctx_projector(image_proj_stage_config, image_proj_model_trainable) def _init_img_ctx_projector(self, config, trainable): self.image_proj_model = instantiate_from_config(config) if not trainable: self.image_proj_model.eval() self.image_proj_model.train = disabled_train for param in self.image_proj_model.parameters(): param.requires_grad = False def _init_embedder(self, config, freeze=True): self.embedder = instantiate_from_config(config) if freeze: self.embedder.eval() self.embedder.train = disabled_train for param in self.embedder.parameters(): param.requires_grad = False def shared_step(self, batch, random_uncond, **kwargs): x, c, fs = self.get_batch_input(batch, random_uncond=random_uncond, return_fs=True) kwargs.update({"fs": fs.long()}) loss, loss_dict = self(x, c, **kwargs) return loss, loss_dict def get_batch_input(self, batch, random_uncond, return_first_stage_outputs=False, return_original_cond=False, return_fs=False, return_cond_frame=False, return_original_input=False, **kwargs): ## x: b c t h w x = super().get_input(batch, self.first_stage_key) ## encode video frames x to z via a 2D encoder z = self.encode_first_stage(x) ## get caption condition cond_input = batch[self.cond_stage_key] if isinstance(cond_input, dict) or isinstance(cond_input, list): cond_emb = self.get_learned_conditioning(cond_input) else: cond_emb = self.get_learned_conditioning(cond_input.to(self.device)) cond = {} ## to support classifier-free guidance, randomly drop out only text conditioning 5%, only image conditioning 5%, and both 5%. if random_uncond: random_num = torch.rand(x.size(0), device=x.device) else: random_num = torch.ones(x.size(0), device=x.device) ## by doning so, we can get text embedding and complete img emb for inference prompt_mask = rearrange(random_num < 2 * self.uncond_prob, "n -> n 1 1") input_mask = 1 - rearrange((random_num >= self.uncond_prob).float() * (random_num < 3 * self.uncond_prob).float(), "n -> n 1 1 1") null_prompt = self.get_learned_conditioning([""]) prompt_imb = torch.where(prompt_mask, null_prompt, cond_emb.detach()) ## get conditioning frame cond_frame_index = 0 if self.rand_cond_frame: cond_frame_index = random.randint(0, self.model.diffusion_model.temporal_length-1) img = x[:,:,cond_frame_index,...] img = input_mask * img ## img: b c h w img_emb = self.embedder(img) ## b l c img_emb = self.image_proj_model(img_emb) if self.model.conditioning_key == 'hybrid': ## simply repeat the cond_frame to match the seq_len of z img_cat_cond = z[:,:,cond_frame_index,:,:] img_cat_cond = img_cat_cond.unsqueeze(2) img_cat_cond = repeat(img_cat_cond, 'b c t h w -> b c (repeat t) h w', repeat=z.shape[2]) cond["c_concat"] = [img_cat_cond] # b c t h w cond["c_crossattn"] = [torch.cat([prompt_imb, img_emb], dim=1)] ## concat in the seq_len dim out = [z, cond] if return_first_stage_outputs: xrec = self.decode_first_stage(z) out.extend([xrec]) if return_original_cond: out.append(cond_input) if return_fs: if self.fps_condition_type == 'fs': fs = super().get_input(batch, 'frame_stride') elif self.fps_condition_type == 'fps': fs = super().get_input(batch, 'fps') out.append(fs) if return_cond_frame: out.append(x[:,:,cond_frame_index,...].unsqueeze(2)) if return_original_input: out.append(x) return out @torch.no_grad() def log_images(self, batch, sample=True, ddim_steps=50, ddim_eta=1., plot_denoise_rows=False, \ unconditional_guidance_scale=1.0, mask=None, **kwargs): """ log images for LatentVisualDiffusion """ ##### sampled_img_num: control sampled imgae for logging, larger value may cause OOM sampled_img_num = 1 for key in batch.keys(): batch[key] = batch[key][:sampled_img_num] ## TBD: currently, classifier_free_guidance sampling is only supported by DDIM use_ddim = ddim_steps is not None log = dict() z, c, xrec, xc, fs, cond_x = self.get_batch_input(batch, random_uncond=False, return_first_stage_outputs=True, return_original_cond=True, return_fs=True, return_cond_frame=True) N = xrec.shape[0] log["image_condition"] = cond_x log["reconst"] = xrec xc_with_fs = [] for idx, content in enumerate(xc): xc_with_fs.append(content + '_fs=' + str(fs[idx].item())) log["condition"] = xc_with_fs kwargs.update({"fs": fs.long()}) c_cat = None if sample: # get uncond embedding for classifier-free guidance sampling if unconditional_guidance_scale != 1.0: if isinstance(c, dict): c_emb = c["c_crossattn"][0] if 'c_concat' in c.keys(): c_cat = c["c_concat"][0] else: c_emb = c if self.uncond_type == "empty_seq": prompts = N * [""] uc_prompt = self.get_learned_conditioning(prompts) elif self.uncond_type == "zero_embed": uc_prompt = torch.zeros_like(c_emb) img = torch.zeros_like(xrec[:,:,0]) ## b c h w ## img: b c h w img_emb = self.embedder(img) ## b l c uc_img = self.image_proj_model(img_emb) uc = torch.cat([uc_prompt, uc_img], dim=1) ## hybrid case if isinstance(c, dict): uc_hybrid = {"c_concat": [c_cat], "c_crossattn": [uc]} uc = uc_hybrid else: uc = None 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, unconditional_guidance_scale=unconditional_guidance_scale, unconditional_conditioning=uc, x0=z, **kwargs) x_samples = self.decode_first_stage(samples) log["samples"] = x_samples if plot_denoise_rows: denoise_grid = self._get_denoise_row_from_list(z_denoise_row) log["denoise_row"] = denoise_grid return log def configure_optimizers(self): """ configure_optimizers for LatentDiffusion """ lr = self.learning_rate params = list(self.model.parameters()) mainlogger.info(f"@Training [{len(params)}] Full Paramters.") if self.cond_stage_trainable: params_cond_stage = [p for p in self.cond_stage_model.parameters() if p.requires_grad == True] mainlogger.info(f"@Training [{len(params_cond_stage)}] Paramters for Cond_stage_model.") params.extend(params_cond_stage) if self.image_proj_model_trainable: mainlogger.info(f"@Training [{len(list(self.image_proj_model.parameters()))}] Paramters for Image_proj_model.") params.extend(list(self.image_proj_model.parameters())) if self.learn_logvar: mainlogger.info('Diffusion model optimizing logvar') if isinstance(params[0], dict): params.append({"params": [self.logvar]}) else: params.append(self.logvar) ## optimizer optimizer = torch.optim.AdamW(params, lr=lr) ## lr scheduler if self.use_scheduler: mainlogger.info("Setting up scheduler...") lr_scheduler = self.configure_schedulers(optimizer) return [optimizer], [lr_scheduler] return optimizer 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 def forward(self, x, t, c_concat: list = None, c_crossattn: list = None, c_adm=None, s=None, mask=None, **kwargs): # temporal_context = fps is foNone 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, **kwargs) elif self.conditioning_key == 'crossattn': cc = torch.cat(c_crossattn, 1) out = self.diffusion_model(x, t, context=cc, **kwargs) elif self.conditioning_key == 'hybrid': ## it is just right [b,c,t,h,w]: concatenate in channel dim xc = torch.cat([x] + c_concat, dim=1) cc = torch.cat(c_crossattn, 1) out = self.diffusion_model(xc, t, context=cc, **kwargs) elif self.conditioning_key == 'resblockcond': cc = c_crossattn[0] out = self.diffusion_model(x, t, context=cc) elif self.conditioning_key == 'adm': cc = c_crossattn[0] out = self.diffusion_model(x, t, y=cc) elif self.conditioning_key == 'hybrid-adm': assert c_adm is not None xc = torch.cat([x] + c_concat, dim=1) cc = torch.cat(c_crossattn, 1) out = self.diffusion_model(xc, t, context=cc, y=c_adm, **kwargs) elif self.conditioning_key == 'hybrid-time': assert s is not None xc = torch.cat([x] + c_concat, dim=1) cc = torch.cat(c_crossattn, 1) out = self.diffusion_model(xc, t, context=cc, s=s) elif self.conditioning_key == 'concat-time-mask': # assert s is not None xc = torch.cat([x] + c_concat, dim=1) out = self.diffusion_model(xc, t, context=None, s=s, mask=mask) elif self.conditioning_key == 'concat-adm-mask': # assert s is not None if c_concat is not None: xc = torch.cat([x] + c_concat, dim=1) else: xc = x out = self.diffusion_model(xc, t, context=None, y=s, mask=mask) elif self.conditioning_key == 'hybrid-adm-mask': cc = torch.cat(c_crossattn, 1) if c_concat is not None: xc = torch.cat([x] + c_concat, dim=1) else: xc = x out = self.diffusion_model(xc, t, context=cc, y=s, mask=mask) elif self.conditioning_key == 'hybrid-time-adm': # adm means y, e.g., class index # assert s is not None assert c_adm is not None xc = torch.cat([x] + c_concat, dim=1) cc = torch.cat(c_crossattn, 1) out = self.diffusion_model(xc, t, context=cc, s=s, y=c_adm) elif self.conditioning_key == 'crossattn-adm': assert c_adm is not None cc = torch.cat(c_crossattn, 1) out = self.diffusion_model(x, t, context=cc, y=c_adm) else: raise NotImplementedError() return out