Merge branch 'prixt-seamless' into development

README.md

@@ -138,6 +138,13 @@ You may also pass a -v<count> option to generate count variants on the original
 passing the first generated image back into img2img the requested number of times. It generates interesting
 variants.
 
+## Seamless Tiling
+
+The seamless tiling mode causes generated images to seamlessly tile with itself. To use it, add the --seamless option when starting the script which will result in all generated images to tile, or for each dream> prompt as shown here:
+```
+dream> "pond garden with lotus by claude monet" --seamless -s100 -n4
+```
+
 ## GFPGAN and Real-ESRGAN Support
 
 The script also provides the ability to do face restoration and
@@ -400,7 +407,11 @@ repository and associated paper for details and limitations.
 
 # Latest Changes
 
-- v1.13 (3 September 2022)
+- v1.14 (In progress)
+
+  - Add "seamless mode" for circular tiling of image. Generates beautiful effects. ([prixt](https://github.com/prixt))
+
+- v1.13 (3 September 2022
 
   - Support image variations (see [VARIATIONS](VARIATIONS.md) ([Kevin Gibbons](https://github.com/bakkot) and many contributors and reviewers)
   - Supports a Google Colab notebook for a standalone server running on Google hardware [Arturo Mendivil](https://github.com/artmen1516)

ldm/dream/pngwriter.py

@@ -59,6 +59,8 @@ def normalize_prompt(self):
         switches.append(f'-H{opt.height       or t2i.height}')
         switches.append(f'-C{opt.cfg_scale    or t2i.cfg_scale}')
         switches.append(f'-A{opt.sampler_name or t2i.sampler_name}')
+        if opt.seamless or t2i.seamless:
+            switches.append(f'--seamless')
         if opt.init_img:
             switches.append(f'-I{opt.init_img}')
         if opt.fit:

ldm/dream/server.py

@@ -70,7 +70,8 @@ def do_POST(self):
         steps = int(post_data['steps'])
         width = int(post_data['width'])
         height = int(post_data['height'])
-        fit    = 'fit' in post_data
+        fit      = 'fit' in post_data
+        seamless = 'seamless' in post_data
         cfgscale = float(post_data['cfgscale'])
         sampler_name  = post_data['sampler']
         gfpgan_strength = float(post_data['gfpgan_strength']) if gfpgan_model_exists else 0
@@ -164,6 +165,7 @@ def image_progress(sample, step):
                                         gfpgan_strength = gfpgan_strength,
                                         upscale         = upscale,
                                         sampler_name    = sampler_name,
+                                        seamless        = seamless,
                                         step_callback=image_progress,
                                         image_callback=image_done)
             else:
@@ -185,6 +187,7 @@ def image_progress(sample, step):
                                             width      = width,
                                             height     = height,
                                             fit        = fit,
+                                            seamless   = seamless,
                                             gfpgan_strength=gfpgan_strength,
                                             upscale         = upscale,
                                             step_callback=image_progress,

ldm/simplet2i.py

@@ -14,6 +14,7 @@
 from tqdm import tqdm, trange
 from itertools import islice
 from einops import rearrange, repeat
+from torch import nn
 from torchvision.utils import make_grid
 from pytorch_lightning import seed_everything
 from torch import autocast
@@ -109,6 +110,7 @@ class T2I:
         downsampling_factor
         precision
         strength
+        seamless
         embedding_path
 
     The vast majority of these arguments default to reasonable values.
@@ -132,6 +134,7 @@ def __init__(
             precision='autocast',
             full_precision=False,
             strength=0.75,  # default in scripts/img2img.py
+            seamless=False,
             embedding_path=None,
             device_type = 'cuda',
             # just to keep track of this parameter when regenerating prompt
@@ -153,6 +156,7 @@ def __init__(
         self.precision                = precision
         self.full_precision           = True if choose_torch_device() == 'mps' else full_precision
         self.strength                 = strength
+        self.seamless                 = seamless
         self.embedding_path           = embedding_path
         self.device_type              = device_type
         self.model                    = None     # empty for now
@@ -217,6 +221,7 @@ def prompt2image(
             step_callback  =    None,
             width          =    None,
             height         =    None,
+            seamless       =    False,
             # these are specific to img2img
             init_img       =    None,
             fit            =    False,
@@ -240,6 +245,7 @@ def prompt2image(
            width                           // width of image, in multiples of 64 (512)
            height                          // height of image, in multiples of 64 (512)
            cfg_scale                       // how strongly the prompt influences the image (7.5) (must be >1)
+           seamless                        // whether the generated image should tile
            init_img                        // path to an initial image - its dimensions override width and height
            strength                        // strength for noising/unnoising init_img. 0.0 preserves image exactly, 1.0 replaces it completely
            gfpgan_strength                 // strength for GFPGAN. 0.0 preserves image exactly, 1.0 replaces it completely
@@ -268,6 +274,7 @@ def process_image(image,seed):
         steps                 = steps      or self.steps
         width                 = width      or self.width
         height                = height     or self.height
+        seamless              = seamless   or self.seamless
         cfg_scale             = cfg_scale  or self.cfg_scale
         ddim_eta              = ddim_eta   or self.ddim_eta
         iterations            = iterations or self.iterations
@@ -278,6 +285,10 @@ def process_image(image,seed):
         model = (
             self.load_model()
         )  # will instantiate the model or return it from cache
+        for m in model.modules():
+            if isinstance(m, (nn.Conv2d, nn.ConvTranspose2d)):
+                m.padding_mode = 'circular' if seamless else m._orig_padding_mode
+
         assert cfg_scale > 1.0, 'CFG_Scale (-C) must be >1.0'
         assert (
             0.0 <= strength <= 1.0
@@ -603,6 +614,10 @@ def load_model(self):
 
             self._set_sampler()
 
+            for m in self.model.modules():
+                if isinstance(m, (nn.Conv2d, nn.ConvTranspose2d)):
+                    m._orig_padding_mode = m.padding_mode
+
         return self.model
 
     # returns a tensor filled with random numbers from a normal distribution

scripts/dream.py

@@ -62,6 +62,7 @@ def main():
         grid  = opt.grid,
         # this is solely for recreating the prompt
         latent_diffusion_weights=opt.laion400m,
+        seamless=opt.seamless,
         embedding_path=opt.embedding_path,
         device_type=opt.device
     )
@@ -87,6 +88,9 @@ def main():
             print(f'{e}. Aborting.')
             sys.exit(-1)
 
+    if opt.seamless:
+        print(">> changed to seamless tiling mode")
+
     # preload the model
     tic = time.time()
     t2i.load_model()
@@ -418,6 +422,11 @@ def create_argv_parser():
         default='outputs/img-samples',
         help='Directory to save generated images and a log of prompts and seeds. Default: outputs/img-samples',
     )
+    parser.add_argument(
+        '--seamless',
+        action='store_true',
+        help='Change the model to seamless tiling (circular) mode',
+    )
     parser.add_argument(
         '--embedding_path',
         type=str,
@@ -540,6 +549,11 @@ def create_cmd_parser():
         default=None,
         help='Directory to save generated images and a log of prompts and seeds',
     )
+    parser.add_argument(
+        '--seamless',
+        action='store_true',
+        help='Change the model to seamless tiling (circular) mode',
+    )
     parser.add_argument(
         '-i',
         '--individual',

static/dream_web/index.html

@@ -37,6 +37,8 @@ <h2 id="header">Stable Diffusion Dream Server</h2>
 	      <option value="k_euler_a">KEULER_A</option>
               <option value="k_heun">KHEUN</option>
             </select>
+            <input type="checkbox" name="seamless" id="seamless">
+	    <label for="seamless">Seamless circular tiling</label>
             <br>
             <label title="Set to multiple of 64" for="width">Width:</label>
             <select id="width" name="width" value="512">
@@ -64,7 +66,7 @@ <h2 id="header">Stable Diffusion Dream Server</h2>
             <input value="-1" type="number" id="seed" name="seed">
             <button type="button" id="reset-seed">&olarr;</button>
             <input type="checkbox" name="progress_images" id="progress_images">
-	    <label for="progress_images">Display in-progress images (slows down generation):</label>
+	    <label for="progress_images">Display in-progress images (slower)</label>
 	    <button type="button" id="reset-all">Reset to Defaults</button>
 	</div>
 	<div id="img2img">
@@ -74,7 +76,7 @@ <h2 id="header">Stable Diffusion Dream Server</h2>
           <label for="strength">Img2Img Strength:</label>
           <input value="0.75" type="number" id="strength" name="strength" step="0.01" min="0" max="1">
           <input type="checkbox" id="fit" name="fit" checked>
-          <label title="Rescale image to fit within requested width and height" for="fit">Fit to width/height:</label>
+          <label title="Rescale image to fit within requested width and height" for="fit">Fit to width/height</label>
 	</div>
         <div id="gfpgan">
           <label title="Strength of the gfpgan (face fixing) algorithm." for="gfpgan_strength">GPFGAN Strength (0 to disable):</label>