make sure learned variance can work for any number of unets in the decoder, defaults to first unet, as suggested was used in the paper

The default config for clip from openai should be ViT-B/32

README.md

@@ -508,7 +508,7 @@ To use a pretrained OpenAI CLIP, simply import `OpenAIClipAdapter` and pass it i
 import torch
 from dalle2_pytorch import DALLE2, DiffusionPriorNetwork, DiffusionPrior, Unet, Decoder, OpenAIClipAdapter
 
-# openai pretrained clip - defaults to ViT/B-32
+# openai pretrained clip - defaults to ViT-B/32
 
 clip = OpenAIClipAdapter()
 
@@ -1003,6 +1003,7 @@ Once built, images will be saved to the same directory the command is invoked
 - [x] pull logic for training diffusion prior into a class DiffusionPriorTrainer, for eventual script based + CLI based training
 - [x] make sure the cascading ddpm in the repository can be trained unconditionally, offer a one-line CLI tool for training on a folder of images
 - [x] bring in cross-scale embedding from iclr paper https://github.com/lucidrains/vit-pytorch/blob/main/vit_pytorch/crossformer.py#L14
+- [x] cross embed layers for downsampling, as an option
 - [ ] become an expert with unets, cleanup unet code, make it fully configurable, port all learnings over to https://github.com/lucidrains/x-unet (test out unet² in ddpm repo) - consider https://github.com/lucidrains/uformer-pytorch attention-based unet
 - [ ] transcribe code to Jax, which lowers the activation energy for distributed training, given access to TPUs
 - [ ] train on a toy task, offer in colab

dalle2_pytorch/dalle2_pytorch.py

@@ -33,6 +33,10 @@ from rotary_embedding_torch import RotaryEmbedding
 from x_clip import CLIP
 from coca_pytorch import CoCa
 
+# constants
+
+NAT = 1. / math.log(2.)
+
 # helper functions
 
 def exists(val):
@@ -41,9 +45,6 @@ def exists(val):
 def identity(t, *args, **kwargs):
     return t
 
-def is_odd(n):
-    return (n % 2) == 1
-
 def default(val, d):
     if exists(val):
         return val
@@ -94,6 +95,9 @@ def freeze_model_and_make_eval_(model):
 
 # tensor helpers
 
+def log(t, eps = 1e-12):
+    return torch.log(t.clamp(min = eps))
+
 def l2norm(t):
     return F.normalize(t, dim = -1)
 
@@ -300,13 +304,43 @@ def noise_like(shape, device, repeat=False):
     noise = lambda: torch.randn(shape, device=device)
     return repeat_noise() if repeat else noise()
 
+def meanflat(x):
+    return x.mean(dim = tuple(range(1, len(x.shape))))
+
+def normal_kl(mean1, logvar1, mean2, logvar2):
+    return 0.5 * (-1.0 + logvar2 - logvar1 + torch.exp(logvar1 - logvar2) + ((mean1 - mean2) ** 2) * torch.exp(-logvar2))
+
+def approx_standard_normal_cdf(x):
+    return 0.5 * (1.0 + torch.tanh(((2.0 / math.pi) ** 0.5) * (x + 0.044715 * (x ** 3))))
+
+def discretized_gaussian_log_likelihood(x, *, means, log_scales, thres = 0.999):
+    assert x.shape == means.shape == log_scales.shape
+
+    centered_x = x - means
+    inv_stdv = torch.exp(-log_scales)
+    plus_in = inv_stdv * (centered_x + 1. / 255.)
+    cdf_plus = approx_standard_normal_cdf(plus_in)
+    min_in = inv_stdv * (centered_x - 1. / 255.)
+    cdf_min = approx_standard_normal_cdf(min_in)
+    log_cdf_plus = log(cdf_plus)
+    log_one_minus_cdf_min = log(1. - cdf_min)
+    cdf_delta = cdf_plus - cdf_min
+
+    log_probs = torch.where(x < -thres,
+        log_cdf_plus,
+        torch.where(x > thres,
+            log_one_minus_cdf_min,
+            log(cdf_delta)))
+
+    return log_probs
+
 def cosine_beta_schedule(timesteps, s = 0.008):
     """
     cosine schedule
     as proposed in https://openreview.net/forum?id=-NEXDKk8gZ
     """
     steps = timesteps + 1
-    x = torch.linspace(0, timesteps, steps)
+    x = torch.linspace(0, timesteps, steps, dtype = torch.float64)
     alphas_cumprod = torch.cos(((x / timesteps) + s) / (1 + s) * torch.pi * 0.5) ** 2
     alphas_cumprod = alphas_cumprod / alphas_cumprod[0]
     betas = 1 - (alphas_cumprod[1:] / alphas_cumprod[:-1])
@@ -317,21 +351,21 @@ def linear_beta_schedule(timesteps):
     scale = 1000 / timesteps
     beta_start = scale * 0.0001
     beta_end = scale * 0.02
-    return torch.linspace(beta_start, beta_end, timesteps)
+    return torch.linspace(beta_start, beta_end, timesteps, dtype = torch.float64)
 
 
 def quadratic_beta_schedule(timesteps):
     scale = 1000 / timesteps
     beta_start = scale * 0.0001
     beta_end = scale * 0.02
-    return torch.linspace(beta_start**2, beta_end**2, timesteps) ** 2
+    return torch.linspace(beta_start**2, beta_end**2, timesteps, dtype = torch.float64) ** 2
 
 
 def sigmoid_beta_schedule(timesteps):
     scale = 1000 / timesteps
     beta_start = scale * 0.0001
     beta_end = scale * 0.02
-    betas = torch.linspace(-6, 6, timesteps)
+    betas = torch.linspace(-6, 6, timesteps, dtype = torch.float64)
     return torch.sigmoid(betas) * (beta_end - beta_start) + beta_start
 
 
@@ -371,17 +405,21 @@ class BaseGaussianDiffusion(nn.Module):
         self.loss_type = loss_type
         self.loss_fn = loss_fn
 
-        self.register_buffer('betas', betas)
-        self.register_buffer('alphas_cumprod', alphas_cumprod)
-        self.register_buffer('alphas_cumprod_prev', alphas_cumprod_prev)
+        # register buffer helper function to cast double back to float
+
+        register_buffer = lambda name, val: self.register_buffer(name, val.to(torch.float32))
+
+        register_buffer('betas', betas)
+        register_buffer('alphas_cumprod', alphas_cumprod)
+        register_buffer('alphas_cumprod_prev', alphas_cumprod_prev)
 
         # calculations for diffusion q(x_t | x_{t-1}) and others
 
-        self.register_buffer('sqrt_alphas_cumprod', torch.sqrt(alphas_cumprod))
-        self.register_buffer('sqrt_one_minus_alphas_cumprod', torch.sqrt(1. - alphas_cumprod))
-        self.register_buffer('log_one_minus_alphas_cumprod', torch.log(1. - alphas_cumprod))
-        self.register_buffer('sqrt_recip_alphas_cumprod', torch.sqrt(1. / alphas_cumprod))
-        self.register_buffer('sqrt_recipm1_alphas_cumprod', torch.sqrt(1. / alphas_cumprod - 1))
+        register_buffer('sqrt_alphas_cumprod', torch.sqrt(alphas_cumprod))
+        register_buffer('sqrt_one_minus_alphas_cumprod', torch.sqrt(1. - alphas_cumprod))
+        register_buffer('log_one_minus_alphas_cumprod', torch.log(1. - alphas_cumprod))
+        register_buffer('sqrt_recip_alphas_cumprod', torch.sqrt(1. / alphas_cumprod))
+        register_buffer('sqrt_recipm1_alphas_cumprod', torch.sqrt(1. / alphas_cumprod - 1))
 
         # calculations for posterior q(x_{t-1} | x_t, x_0)
 
@@ -389,19 +427,13 @@ class BaseGaussianDiffusion(nn.Module):
 
         # above: equal to 1. / (1. / (1. - alpha_cumprod_tm1) + alpha_t / beta_t)
 
-        self.register_buffer('posterior_variance', posterior_variance)
+        register_buffer('posterior_variance', 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', torch.log(posterior_variance.clamp(min =1e-20)))
-        self.register_buffer('posterior_mean_coef1', betas * torch.sqrt(alphas_cumprod_prev) / (1. - alphas_cumprod))
-        self.register_buffer('posterior_mean_coef2', (1. - alphas_cumprod_prev) * torch.sqrt(alphas) / (1. - alphas_cumprod))
-
-    def q_mean_variance(self, x_start, t):
-        mean = extract(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start
-        variance = extract(1. - self.alphas_cumprod, t, x_start.shape)
-        log_variance = extract(self.log_one_minus_alphas_cumprod, t, x_start.shape)
-        return mean, variance, log_variance
+        register_buffer('posterior_log_variance_clipped', torch.log(posterior_variance.clamp(min =1e-20)))
+        register_buffer('posterior_mean_coef1', betas * torch.sqrt(alphas_cumprod_prev) / (1. - alphas_cumprod))
+        register_buffer('posterior_mean_coef2', (1. - alphas_cumprod_prev) * torch.sqrt(alphas) / (1. - alphas_cumprod))
 
     def q_posterior(self, x_start, x_t, t):
         posterior_mean = (
@@ -830,7 +862,7 @@ class DiffusionPrior(BaseGaussianDiffusion):
         image_channels = 3,
         timesteps = 1000,
         cond_drop_prob = 0.,
-        loss_type = "l1",
+        loss_type = "l2",
         predict_x_start = True,
         beta_schedule = "cosine",
         condition_on_text_encodings = True, # the paper suggests this is needed, but you can turn it off for your CLIP preprocessed text embed -> image embed training
@@ -1235,12 +1267,13 @@ class CrossEmbedLayer(nn.Module):
     def __init__(
         self,
         dim_in,
-        dim_out,
         kernel_sizes,
+        dim_out = None,
         stride = 2
     ):
         super().__init__()
-        assert all([*map(is_odd, kernel_sizes)])
+        assert all([*map(lambda t: (t % 2) == (stride % 2), kernel_sizes)])
+        dim_out = default(dim_out, dim_in)
 
         kernel_sizes = sorted(kernel_sizes)
         num_scales = len(kernel_sizes)
@@ -1270,6 +1303,7 @@ class Unet(nn.Module):
         out_dim = None,
         dim_mults=(1, 2, 4, 8),
         channels = 3,
+        channels_out = None,
         attn_dim_head = 32,
         attn_heads = 16,
         lowres_cond = False, # for cascading diffusion - https://cascaded-diffusion.github.io/
@@ -1282,6 +1316,8 @@ class Unet(nn.Module):
         init_conv_kernel_size = 7,
         resnet_groups = 8,
         init_cross_embed_kernel_sizes = (3, 7, 15),
+        cross_embed_downsample = False,
+        cross_embed_downsample_kernel_sizes = (2, 4),
         **kwargs
     ):
         super().__init__()
@@ -1298,11 +1334,12 @@ class Unet(nn.Module):
         # determine dimensions
 
         self.channels = channels
+        self.channels_out = default(channels_out, channels)
 
         init_channels = channels if not lowres_cond else channels * 2 # in cascading diffusion, one concats the low resolution image, blurred, for conditioning the higher resolution synthesis
         init_dim = default(init_dim, dim // 3 * 2)
 
-        self.init_conv = CrossEmbedLayer(init_channels, init_dim, init_cross_embed_kernel_sizes, stride = 1)
+        self.init_conv = CrossEmbedLayer(init_channels, dim_out = init_dim, kernel_sizes = init_cross_embed_kernel_sizes, stride = 1)
 
         dims = [init_dim, *map(lambda m: dim * m, dim_mults)]
         in_out = list(zip(dims[:-1], dims[1:]))
@@ -1362,6 +1399,12 @@ class Unet(nn.Module):
 
         assert len(resnet_groups) == len(in_out)
 
+        # downsample klass
+
+        downsample_klass = Downsample
+        if cross_embed_downsample:
+            downsample_klass = partial(CrossEmbedLayer, kernel_sizes = cross_embed_downsample_kernel_sizes)
+
         # layers
 
         self.downs = nn.ModuleList([])
@@ -1377,7 +1420,7 @@ class Unet(nn.Module):
                 ResnetBlock(dim_in, dim_out, time_cond_dim = time_cond_dim, groups = groups),
                 Residual(LinearAttention(dim_out, **attn_kwargs)) if sparse_attn else nn.Identity(),
                 ResnetBlock(dim_out, dim_out, cond_dim = layer_cond_dim, time_cond_dim = time_cond_dim, groups = groups),
-                Downsample(dim_out) if not is_last else nn.Identity()
+                downsample_klass(dim_out) if not is_last else nn.Identity()
             ]))
 
         mid_dim = dims[-1]
@@ -1397,11 +1440,9 @@ class Unet(nn.Module):
                 Upsample(dim_in)
             ]))
 
-        out_dim = default(out_dim, channels)
-
         self.final_conv = nn.Sequential(
             ResnetBlock(dim, dim, groups = resnet_groups[0]),
-            nn.Conv2d(dim, out_dim, 1)
+            nn.Conv2d(dim, self.channels_out, 1)
         )
 
     # if the current settings for the unet are not correct
@@ -1411,13 +1452,25 @@ class Unet(nn.Module):
         *,
         lowres_cond,
         channels,
+        channels_out,
         cond_on_image_embeds,
         cond_on_text_encodings
     ):
-        if lowres_cond == self.lowres_cond and channels == self.channels and cond_on_image_embeds == self.cond_on_image_embeds and cond_on_text_encodings == self.cond_on_text_encodings:
+        if lowres_cond == self.lowres_cond and \
+            channels == self.channels and \
+            cond_on_image_embeds == self.cond_on_image_embeds and \
+            cond_on_text_encodings == self.cond_on_text_encodings and \
+            channels_out == self.channels_out:
             return self
 
-        updated_kwargs = {'lowres_cond': lowres_cond, 'channels': channels, 'cond_on_image_embeds': cond_on_image_embeds, 'cond_on_text_encodings': cond_on_text_encodings}
+        updated_kwargs = dict(
+            lowres_cond = lowres_cond,
+            channels = channels,
+            channels_out = channels_out,
+            cond_on_image_embeds = cond_on_image_embeds,
+            cond_on_text_encodings = cond_on_text_encodings
+        )
+
         return self.__class__(**{**self._locals, **updated_kwargs})
 
     def forward_with_cond_scale(
@@ -1482,11 +1535,12 @@ class Unet(nn.Module):
 
         if self.cond_on_image_embeds:
             image_tokens = self.image_to_cond(image_embed)
+            null_image_embed = self.null_image_embed.to(image_tokens.dtype) # for some reason pytorch AMP not working
 
             image_tokens = torch.where(
                 image_keep_mask,
                 image_tokens,
-                self.null_image_embed
+                null_image_embed
             )
 
         # take care of text encodings (optional)
@@ -1510,10 +1564,12 @@ class Unet(nn.Module):
                 text_mask = rearrange(text_mask, 'b n -> b n 1')
                 text_keep_mask = text_mask & text_keep_mask
 
+            null_text_embed = self.null_text_embed.to(text_tokens.dtype) # for some reason pytorch AMP not working
+
             text_tokens = torch.where(
                 text_keep_mask,
                 text_tokens,
-                self.null_text_embed
+                null_text_embed
             )
 
         # main conditioning tokens (c)
@@ -1601,7 +1657,7 @@ class Decoder(BaseGaussianDiffusion):
         timesteps = 1000,
         image_cond_drop_prob = 0.1,
         text_cond_drop_prob = 0.5,
-        loss_type = 'l1',
+        loss_type = 'l2',
         beta_schedule = 'cosine',
         predict_x_start = False,
         predict_x_start_for_latent_diffusion = False,
@@ -1614,6 +1670,8 @@ class Decoder(BaseGaussianDiffusion):
         clip_denoised = True,
         clip_x_start = True,
         clip_adapter_overrides = dict(),
+        learned_variance = True,
+        vb_loss_weight = 0.001,
         unconditional = False
     ):
         super().__init__(
@@ -1652,10 +1710,18 @@ class Decoder(BaseGaussianDiffusion):
         unets = cast_tuple(unet)
         vaes = pad_tuple_to_length(cast_tuple(vae), len(unets), fillvalue = NullVQGanVAE(channels = self.channels))
 
+        # whether to use learned variance, defaults to True for the first unet in the cascade, as in paper
+
+        learned_variance = pad_tuple_to_length(cast_tuple(learned_variance), len(unets), fillvalue = False)
+        self.learned_variance = learned_variance
+        self.vb_loss_weight = vb_loss_weight
+
+        # construct unets and vaes
+
         self.unets = nn.ModuleList([])
         self.vaes = nn.ModuleList([])
 
-        for ind, (one_unet, one_vae) in enumerate(zip(unets, vaes)):
+        for ind, (one_unet, one_vae, one_unet_learned_var) in enumerate(zip(unets, vaes, learned_variance)):
             assert isinstance(one_unet, Unet)
             assert isinstance(one_vae, (VQGanVAE, NullVQGanVAE))
 
@@ -1663,12 +1729,14 @@ class Decoder(BaseGaussianDiffusion):
             latent_dim = one_vae.encoded_dim if exists(one_vae) else None
 
             unet_channels = default(latent_dim, self.channels)
+            unet_channels_out = unet_channels * (1 if not one_unet_learned_var else 2)
 
             one_unet = one_unet.cast_model_parameters(
                 lowres_cond = not is_first,
                 cond_on_image_embeds = is_first and not unconditional,
                 cond_on_text_encodings = one_unet.cond_on_text_encodings and not unconditional,
-                channels = unet_channels
+                channels = unet_channels,
+                channels_out = unet_channels_out
             )
 
             self.unets.append(one_unet)
@@ -1731,8 +1799,11 @@ class Decoder(BaseGaussianDiffusion):
         yield
         unet.cpu()
 
-    def p_mean_variance(self, unet, x, t, image_embed, text_encodings = None, text_mask = None, lowres_cond_img = None, clip_denoised = True, predict_x_start = False, cond_scale = 1.):
-        pred = unet.forward_with_cond_scale(x, t, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, cond_scale = cond_scale, lowres_cond_img = lowres_cond_img)
+    def p_mean_variance(self, unet, x, t, image_embed, text_encodings = None, text_mask = None, lowres_cond_img = None, clip_denoised = True, predict_x_start = False, learned_variance = False, cond_scale = 1., model_output = None):
+        pred = default(model_output, lambda: unet.forward_with_cond_scale(x, t, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, cond_scale = cond_scale, lowres_cond_img = lowres_cond_img))
+
+        if learned_variance:
+            pred, var_interp_frac_unnormalized = pred.chunk(2, dim = 1)
 
         if predict_x_start:
             x_recon = pred
@@ -1743,19 +1814,31 @@ class Decoder(BaseGaussianDiffusion):
             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 learned_variance:
+            # if learned variance, posterio variance and posterior log variance are predicted by the network
+            # by an interpolation of the max and min log beta values
+            # eq 15 - https://arxiv.org/abs/2102.09672
+            min_log = extract(self.posterior_log_variance_clipped, t, x.shape)
+            max_log = extract(torch.log(self.betas), t, x.shape)
+            var_interp_frac = unnormalize_img(var_interp_frac_unnormalized)
+
+            posterior_log_variance = var_interp_frac * max_log + (1 - var_interp_frac) * min_log
+            posterior_variance = posterior_log_variance.exp()
+
         return model_mean, posterior_variance, posterior_log_variance
 
     @torch.inference_mode()
-    def p_sample(self, unet, x, t, image_embed, text_encodings = None, text_mask = None, cond_scale = 1., lowres_cond_img = None, predict_x_start = False, clip_denoised = True, repeat_noise = False):
+    def p_sample(self, unet, x, t, image_embed, text_encodings = None, text_mask = None, cond_scale = 1., lowres_cond_img = None, predict_x_start = False, learned_variance = False, clip_denoised = True, repeat_noise = False):
         b, *_, device = *x.shape, x.device
-        model_mean, _, model_log_variance = self.p_mean_variance(unet, x = x, t = t, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, cond_scale = cond_scale, lowres_cond_img = lowres_cond_img, clip_denoised = clip_denoised, predict_x_start = predict_x_start)
+        model_mean, _, model_log_variance = self.p_mean_variance(unet, x = x, t = t, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, cond_scale = cond_scale, lowres_cond_img = lowres_cond_img, clip_denoised = clip_denoised, predict_x_start = predict_x_start, learned_variance = learned_variance)
         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.inference_mode()
-    def p_sample_loop(self, unet, shape, image_embed, predict_x_start = False, clip_denoised = True, lowres_cond_img = None, text_encodings = None, text_mask = None, cond_scale = 1):
+    def p_sample_loop(self, unet, shape, image_embed, predict_x_start = False, learned_variance = False, clip_denoised = True, lowres_cond_img = None, text_encodings = None, text_mask = None, cond_scale = 1):
         device = self.betas.device
 
         b = shape[0]
@@ -1772,17 +1855,18 @@ class Decoder(BaseGaussianDiffusion):
                 cond_scale = cond_scale,
                 lowres_cond_img = lowres_cond_img,
                 predict_x_start = predict_x_start,
+                learned_variance = learned_variance,
                 clip_denoised = clip_denoised
             )
 
         return img
 
-    def p_losses(self, unet, x_start, times, *, image_embed, lowres_cond_img = None, text_encodings = None, text_mask = None, predict_x_start = False, noise = None):
+    def p_losses(self, unet, x_start, times, *, image_embed, lowres_cond_img = None, text_encodings = None, text_mask = None, predict_x_start = False, noise = None, learned_variance = False, clip_denoised = False):
         noise = default(noise, lambda: torch.randn_like(x_start))
 
         x_noisy = self.q_sample(x_start = x_start, t = times, noise = noise)
 
-        pred = unet(
+        model_output = unet(
             x_noisy,
             times,
             image_embed = image_embed,
@@ -1793,10 +1877,43 @@ class Decoder(BaseGaussianDiffusion):
             text_cond_drop_prob = self.text_cond_drop_prob,
         )
 
+        if learned_variance:
+            pred, _ = model_output.chunk(2, dim = 1)
+        else:
+            pred = model_output
+
         target = noise if not predict_x_start else x_start
 
         loss = self.loss_fn(pred, target)
-        return loss
+
+        if not learned_variance:
+            # return simple loss if not using learned variance
+            return loss
+
+        # if learning the variance, also include the extra weight kl loss
+
+        true_mean, _, true_log_variance_clipped = self.q_posterior(x_start = x_start, x_t = x_noisy, t = times)
+        model_mean, _, model_log_variance = self.p_mean_variance(unet, x = x_noisy, t = times, image_embed = image_embed, clip_denoised = clip_denoised, learned_variance = True, model_output = model_output)
+
+        # kl loss with detached model predicted mean, for stability reasons as in paper
+
+        detached_model_mean = model_mean.detach()
+
+        kl = normal_kl(true_mean, true_log_variance_clipped, detached_model_mean, model_log_variance)
+        kl = meanflat(kl) * NAT
+
+        decoder_nll = -discretized_gaussian_log_likelihood(x_start, means = detached_model_mean, log_scales = 0.5 * model_log_variance)
+        decoder_nll = meanflat(decoder_nll) * NAT
+
+        # at the first timestep return the decoder NLL, otherwise return KL(q(x_{t-1}|x_t,x_0) || p(x_{t-1}|x_t))
+
+        vb_losses = torch.where(times == 0, decoder_nll, kl)
+
+        # weight the vb loss smaller, for stability, as in the paper (recommended 0.001)
+
+        vb_loss = vb_losses.mean() * self.vb_loss_weight
+
+        return loss + vb_loss
 
     @torch.inference_mode()
     @eval_decorator
@@ -1823,7 +1940,7 @@ class Decoder(BaseGaussianDiffusion):
         img = None
         is_cuda = next(self.parameters()).is_cuda
 
-        for unet_number, unet, vae, channel, image_size, predict_x_start in tqdm(zip(range(1, len(self.unets) + 1), self.unets, self.vaes, self.sample_channels, self.image_sizes, self.predict_x_start)):
+        for unet_number, unet, vae, channel, image_size, predict_x_start, learned_variance in tqdm(zip(range(1, len(self.unets) + 1), self.unets, self.vaes, self.sample_channels, self.image_sizes, self.predict_x_start, self.learned_variance)):
 
             context = self.one_unet_in_gpu(unet = unet) if is_cuda else null_context()
 
@@ -1849,6 +1966,7 @@ class Decoder(BaseGaussianDiffusion):
                     text_mask = text_mask,
                     cond_scale = cond_scale,
                     predict_x_start = predict_x_start,
+                    learned_variance = learned_variance,
                     clip_denoised = not is_latent_diffusion,
                     lowres_cond_img = lowres_cond_img
                 )
@@ -1878,6 +1996,7 @@ class Decoder(BaseGaussianDiffusion):
         target_image_size   = self.image_sizes[unet_index]
         predict_x_start     = self.predict_x_start[unet_index]
         random_crop_size    = self.random_crop_sizes[unet_index]
+        learned_variance    = self.learned_variance[unet_index]
         b, c, h, w, device, = *image.shape, image.device
 
         check_shape(image, 'b c h w', c = self.channels)
@@ -1915,7 +2034,7 @@ class Decoder(BaseGaussianDiffusion):
             if exists(lowres_cond_img):
                 lowres_cond_img = vae.encode(lowres_cond_img)
 
-        return self.p_losses(unet, image, times, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, lowres_cond_img = lowres_cond_img, predict_x_start = predict_x_start)
+        return self.p_losses(unet, image, times, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, lowres_cond_img = lowres_cond_img, predict_x_start = predict_x_start, learned_variance = learned_variance)
 
 # main class
 
@@ -1965,4 +2084,3 @@ class DALLE2(nn.Module):
             return images[0]
 
         return images
-

dalle2_pytorch/train.py

@@ -105,6 +105,10 @@ class EMA(nn.Module):
         self.register_buffer('initted', torch.Tensor([False]))
         self.register_buffer('step', torch.tensor([0.]))
 
+    def restore_ema_model_device(self):
+        device = self.initted.device
+        self.ema_model.to(device)
+
     def update(self):
         self.step += 1
 
@@ -305,6 +309,11 @@ class DecoderTrainer(nn.Module):
 
         if self.use_ema:
             self.decoder.unets = trainable_unets             # restore original training unets
+
+        # cast the ema_model unets back to original device
+        for ema in self.ema_unets:
+            ema.restore_ema_model_device()
+
         return output
 
     def forward(

setup.py

@@ -10,7 +10,7 @@ setup(
       'dream = dalle2_pytorch.cli:dream'
     ],
   },
-  version = '0.2.7',
+  version = '0.2.14',
   license='MIT',
   description = 'DALL-E 2',
   author = 'Phil Wang',