normalize conditioning tokens outside of cross attention blocks

README.md

@@ -1017,6 +1017,7 @@ Once built, images will be saved to the same directory the command is invoked
 - [ ] offer save / load methods on the trainer classes to automatically take care of state dicts for scalers / optimizers / saving versions and checking for breaking changes
 - [ ] bring in skip-layer excitatons (from lightweight gan paper) to see if it helps for either decoder of unet or vqgan-vae training
 - [ ] decoder needs one day worth of refactor for tech debt
+- [ ] allow for unet to be able to condition non-cross attention style as well
 
 ## Citations
 

dalle2_pytorch/dalle2_pytorch.py

@@ -1,7 +1,7 @@
 import math
 from tqdm import tqdm
 from inspect import isfunction
-from functools import partial
+from functools import partial, wraps
 from contextlib import contextmanager
 from collections import namedtuple
 from pathlib import Path
@@ -45,6 +45,14 @@ def exists(val):
 def identity(t, *args, **kwargs):
     return t
 
+def maybe(fn):
+    @wraps(fn)
+    def inner(x):
+        if not exists(x):
+            return x
+        return fn(x)
+    return inner
+
 def default(val, d):
     if exists(val):
         return val
@@ -606,7 +614,6 @@ class Attention(nn.Module):
         heads = 8,
         dropout = 0.,
         causal = False,
-        post_norm = False,
         rotary_emb = None
     ):
         super().__init__()
@@ -616,7 +623,6 @@ class Attention(nn.Module):
 
         self.causal = causal
         self.norm = LayerNorm(dim)
-        self.post_norm = LayerNorm(dim)     # sandwich norm from Coqview paper + Normformer
         self.dropout = nn.Dropout(dropout)
 
         self.null_kv = nn.Parameter(torch.randn(2, dim_head))
@@ -627,7 +633,7 @@ class Attention(nn.Module):
 
         self.to_out = nn.Sequential(
             nn.Linear(inner_dim, dim, bias = False),
-            LayerNorm(dim) if post_norm else nn.Identity()
+            LayerNorm(dim)
         )
 
     def forward(self, x, mask = None, attn_bias = None):
@@ -684,8 +690,7 @@ class Attention(nn.Module):
         out = einsum('b h i j, b j d -> b h i d', attn, v)
 
         out = rearrange(out, 'b h n d -> b n (h d)')
-        out = self.to_out(out)
-        return self.post_norm(out)
+        return self.to_out(out)
 
 class CausalTransformer(nn.Module):
     def __init__(
@@ -711,7 +716,7 @@ class CausalTransformer(nn.Module):
         self.layers = nn.ModuleList([])
         for _ in range(depth):
             self.layers.append(nn.ModuleList([
-                Attention(dim = dim, causal = True, dim_head = dim_head, heads = heads, dropout = attn_dropout, post_norm = normformer, rotary_emb = rotary_emb),
+                Attention(dim = dim, causal = True, dim_head = dim_head, heads = heads, dropout = attn_dropout, rotary_emb = rotary_emb),
                 FeedForward(dim = dim, mult = ff_mult, dropout = ff_dropout, post_activation_norm = normformer)
             ]))
 
@@ -1158,6 +1163,7 @@ class CrossAttention(nn.Module):
         dim_head = 64,
         heads = 8,
         dropout = 0.,
+        norm_context = False
     ):
         super().__init__()
         self.scale = dim_head ** -0.5
@@ -1167,13 +1173,17 @@ class CrossAttention(nn.Module):
         context_dim = default(context_dim, dim)
 
         self.norm = LayerNorm(dim)
-        self.norm_context = LayerNorm(context_dim)
+        self.norm_context = LayerNorm(context_dim) if norm_context else nn.Identity()
         self.dropout = nn.Dropout(dropout)
 
         self.null_kv = nn.Parameter(torch.randn(2, dim_head))
         self.to_q = nn.Linear(dim, inner_dim, bias = False)
         self.to_kv = nn.Linear(context_dim, inner_dim * 2, bias = False)
-        self.to_out = nn.Linear(inner_dim, dim, bias = False)
+
+        self.to_out = nn.Sequential(
+            nn.Linear(inner_dim, dim, bias = False),
+            LayerNorm(dim)
+        )
 
     def forward(self, x, context, mask = None):
         b, n, device = *x.shape[:2], x.device
@@ -1369,6 +1379,9 @@ class Unet(nn.Module):
             Rearrange('b (n d) -> b n d', n = num_image_tokens)
         ) if image_embed_dim != cond_dim else nn.Identity()
 
+        self.norm_cond = nn.LayerNorm(cond_dim)
+        self.norm_mid_cond = nn.LayerNorm(cond_dim)
+
         # text encoding conditioning (optional)
 
         self.text_to_cond = None
@@ -1584,6 +1597,11 @@ class Unet(nn.Module):
 
         mid_c = c if not exists(text_tokens) else torch.cat((c, text_tokens), dim = -2)
 
+        # normalize conditioning tokens
+
+        c = self.norm_cond(c)
+        mid_c = self.norm_mid_cond(mid_c)
+
         # go through the layers of the unet, down and up
 
         hiddens = []
@@ -1844,6 +1862,8 @@ class Decoder(BaseGaussianDiffusion):
         b = shape[0]
         img = torch.randn(shape, device = device)
 
+        lowres_cond_img = maybe(normalize_neg_one_to_one)(lowres_cond_img)
+
         for i in tqdm(reversed(range(0, self.num_timesteps)), desc = 'sampling loop time step', total = self.num_timesteps):
             img = self.p_sample(
                 unet,
@@ -1868,9 +1888,7 @@ class Decoder(BaseGaussianDiffusion):
         # normalize to [-1, 1]
 
         x_start = normalize_neg_one_to_one(x_start)
-
-        if exists(lowres_cond_img):
-            lowres_cond_img = normalize_neg_one_to_one(lowres_cond_img)
+        lowres_cond_img = maybe(normalize_neg_one_to_one)(lowres_cond_img)
 
         # get x_t
 

dalle2_pytorch/optimizer.py

@@ -7,16 +7,17 @@ def separate_weight_decayable_params(params):
 
 def get_optimizer(
     params,
-    lr = 3e-4,
+    lr = 2e-5,
     wd = 1e-2,
     betas = (0.9, 0.999),
+    eps = 1e-8,
     filter_by_requires_grad = False
 ):
     if filter_by_requires_grad:
         params = list(filter(lambda t: t.requires_grad, params))
 
     if wd == 0:
-        return Adam(params, lr = lr, betas = betas)
+        return Adam(params, lr = lr, betas = betas, eps = eps)
 
     params = set(params)
     wd_params, no_wd_params = separate_weight_decayable_params(params)
@@ -26,4 +27,4 @@ def get_optimizer(
         {'params': list(no_wd_params), 'weight_decay': 0},
     ]
 
-    return AdamW(param_groups, lr = lr, weight_decay = wd, betas = betas)
+    return AdamW(param_groups, lr = lr, weight_decay = wd, betas = betas, eps = eps)

dalle2_pytorch/train.py

@@ -90,7 +90,7 @@ class EMA(nn.Module):
     def __init__(
         self,
         model,
-        beta = 0.99,
+        beta = 0.9999,
         update_after_step = 1000,
         update_every = 10,
     ):
@@ -147,6 +147,7 @@ class DiffusionPriorTrainer(nn.Module):
         use_ema = True,
         lr = 3e-4,
         wd = 1e-2,
+        eps = 1e-6,
         max_grad_norm = None,
         amp = False,
         **kwargs
@@ -173,6 +174,7 @@ class DiffusionPriorTrainer(nn.Module):
             diffusion_prior.parameters(),
             lr = lr,
             wd = wd,
+            eps = eps,
             **kwargs
         )
 
@@ -221,8 +223,9 @@ class DecoderTrainer(nn.Module):
         self,
         decoder,
         use_ema = True,
-        lr = 3e-4,
+        lr = 2e-5,
         wd = 1e-2,
+        eps = 1e-8,
         max_grad_norm = None,
         amp = False,
         **kwargs
@@ -247,13 +250,14 @@ class DecoderTrainer(nn.Module):
         # be able to finely customize learning rate, weight decay
         # per unet
 
-        lr, wd = map(partial(cast_tuple, length = self.num_unets), (lr, wd))
+        lr, wd, eps = map(partial(cast_tuple, length = self.num_unets), (lr, wd, eps))
 
-        for ind, (unet, unet_lr, unet_wd) in enumerate(zip(self.decoder.unets, lr, wd)):
+        for ind, (unet, unet_lr, unet_wd, unet_eps) in enumerate(zip(self.decoder.unets, lr, wd, eps)):
             optimizer = get_optimizer(
                 unet.parameters(),
                 lr = unet_lr,
                 wd = unet_wd,
+                eps = unet_eps,
                 **kwargs
             )
 

setup.py

@@ -10,7 +10,7 @@ setup(
       'dream = dalle2_pytorch.cli:dream'
     ],
   },
-  version = '0.2.16',
+  version = '0.2.23',
   license='MIT',
   description = 'DALL-E 2',
   author = 'Phil Wang',