Merge branch 'asymmetric-tiling' of https://github.com/carson-katri/InvokeAI into carson-katri-asymmetric-tiling

docs/features/CLI.md

@@ -159,6 +159,7 @@ Here are the invoke> command that apply to txt2img:
 | --individual       | -i        | True                | Turn off grid mode (deprecated; leave off --grid instead) |
 | --outdir <path>    |  -o<path> | outputs/img_samples  | Temporarily change the location of these images |
 | --seamless         |           | False               | Activate seamless tiling for interesting effects |
+| --seamless_axes    |           | x,y                 | Specify which axes to use circular convolution on. |
 | --log_tokenization | -t        | False               | Display a color-coded list of the parsed tokens derived from the prompt |
 | --skip_normalization| -x       | False               | Weighted subprompts will not be normalized. See [Weighted Prompts](./OTHER.md#weighted-prompts) |
 | --upscale <int> <float> | -U <int> <float> | -U 1 0.75| Upscale image by magnification factor (2, 4), and set strength of upscaling (0.0-1.0). If strength not set, will default to 0.75. |

docs/features/OTHER.md

@@ -26,6 +26,12 @@ for each `invoke>` prompt as shown here:
 invoke> "pond garden with lotus by claude monet" --seamless -s100 -n4
 ```
 
+By default this will tile on both the X and Y axes. However, you can also specify specific axes to tile on with `--seamless_axes`.
+Possible values are `x`, `y`, and `x,y`:
+```python
+invoke> "pond garden with lotus by claude monet" --seamless --seamless_axes=x -s100 -n4
+```
+
 ---
 
 ## **Shortcuts: Reusing Seeds**

ldm/generate.py

@@ -34,6 +34,7 @@ from ldm.invoke.image_util import InitImageResizer
 from ldm.invoke.devices import choose_torch_device, choose_precision
 from ldm.invoke.conditioning import get_uc_and_c
 from ldm.invoke.model_cache import ModelCache
+from ldm.invoke.seamless import configure_model_padding
 from ldm.invoke.txt2mask import Txt2Mask, SegmentedGrayscale
     
 def fix_func(orig):
@@ -174,6 +175,7 @@ class Generate:
         self.precision      = precision
         self.strength       = 0.75
         self.seamless       = False
+        self.seamless_axes  = {'x','y'}
         self.hires_fix      = False
         self.embedding_path = embedding_path
         self.model          = None     # empty for now
@@ -260,6 +262,7 @@ class Generate:
             height           = None,
             sampler_name     = None,
             seamless         = False,
+            seamless_axes    = {'x','y'},
             log_tokenization = False,
             with_variations  = None,
             variation_amount = 0.0,
@@ -335,6 +338,7 @@ class Generate:
         width = width or self.width
         height = height or self.height
         seamless = seamless or self.seamless
+        seamless_axes = seamless_axes or self.seamless_axes
         hires_fix = hires_fix or self.hires_fix
         cfg_scale = cfg_scale or self.cfg_scale
         ddim_eta = ddim_eta or self.ddim_eta
@@ -352,10 +356,8 @@ class Generate:
         # to the width and height of the image training set
         width = width or self.width
         height = height or self.height
-        
-        for m in model.modules():
-            if isinstance(m, (nn.Conv2d, nn.ConvTranspose2d)):
-                m.padding_mode = 'circular' if seamless else m._orig_padding_mode
+
+        configure_model_padding(model, seamless, seamless_axes)
 
         assert cfg_scale > 1.0, 'CFG_Scale (-C) must be >1.0'
         assert threshold >= 0.0, '--threshold must be >=0.0'

ldm/invoke/args.py

@@ -784,6 +784,12 @@ class Args(object):
             action='store_true',
             help='Change the model to seamless tiling (circular) mode',
         )
+        special_effects_group.add_argument(
+            '--seamless_axes',
+            default=['x', 'y'],
+            type=list[str],
+            help='Specify which axes to use circular convolution on.',
+        )
         variation_group.add_argument(
             '-v',
             '--variation_amount',

ldm/invoke/seamless.py

@@ -0,0 +1,30 @@
+import torch.nn as nn
+        
+def _conv_forward_asymmetric(self, input, weight, bias):
+    """
+    Patch for Conv2d._conv_forward that supports asymmetric padding
+    """
+    working = nn.functional.pad(input, self.asymmetric_padding['x'], mode=self.asymmetric_padding_mode['x'])
+    working = nn.functional.pad(working, self.asymmetric_padding['y'], mode=self.asymmetric_padding_mode['y'])
+    return nn.functional.conv2d(working, weight, bias, self.stride, nn.modules.utils._pair(0), self.dilation, self.groups)
+
+def configure_model_padding(model, seamless, seamless_axes):
+    """
+    Modifies the 2D convolution layers to use a circular padding mode based on the `seamless` and `seamless_axes` options.
+    """
+    for m in model.modules():
+        if isinstance(m, (nn.Conv2d, nn.ConvTranspose2d)):
+            if seamless:
+                m.asymmetric_padding_mode = {}
+                m.asymmetric_padding = {}
+                m.asymmetric_padding_mode['x'] = 'circular' if ('x' in seamless_axes) else 'constant'
+                m.asymmetric_padding['x'] = (m._reversed_padding_repeated_twice[0], m._reversed_padding_repeated_twice[1], 0, 0)
+                m.asymmetric_padding_mode['y'] = 'circular' if ('y' in seamless_axes) else 'constant'
+                m.asymmetric_padding['y'] = (0, 0, m._reversed_padding_repeated_twice[2], m._reversed_padding_repeated_twice[3])
+                m._conv_forward = _conv_forward_asymmetric.__get__(m, nn.Conv2d)
+            else:
+                m._conv_forward = nn.Conv2d._conv_forward.__get__(m, nn.Conv2d)
+                if hasattr(m, 'asymmetric_padding_mode'):
+                    del m.asymmetric_padding_mode
+                if hasattr(m, 'asymmetric_padding'):
+                    del m.asymmetric_padding