fix issues

environment/environment.yaml

@@ -0,0 +1,30 @@
+name: ldm
+channels:
+  - pytorch
+  - defaults
+dependencies:
+  - python=3.8.5
+  - pip=20.3
+  - cudatoolkit=11.3
+  - pytorch=1.11.0
+  - torchvision=0.12.0
+  - numpy=1.19.2
+  - pip:
+    - albumentations==0.4.3
+    - diffusers
+    - opencv-python==4.1.2.30
+    - pudb==2019.2
+    - invisible-watermark
+    - imageio==2.9.0
+    - imageio-ffmpeg==0.4.2
+    - pytorch-lightning==1.4.2
+    - omegaconf==2.1.1
+    - test-tube>=0.7.5
+    - streamlit>=0.73.1
+    - einops==0.3.0
+    - torch-fidelity==0.3.0
+    - torchmetrics==0.6.0
+    - kornia==0.6
+    - -e git+https://github.com/CompVis/taming-transformers.git@master#egg=taming-transformers
+    - -e git+https://github.com/openai/CLIP.git@main#egg=clip
+    - -r requirements.txt

environment/requirements.txt

@@ -0,0 +1,10 @@
+ftfy
+opencv-python
+ipywidgets
+matplotlib
+pyrallis
+torch==1.12.0
+diffusers==0.12.1
+transformers==4.26.0
+jupyter
+accelerate

pipeline_attend_and_excite.py

@@ -17,7 +17,6 @@
 from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput
 from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
-#from pipeline_sd import StableDiffusionPipeline
 from diffusers.pipelines.stable_diffusion import StableDiffusionPipeline
 
 from utils.gaussian_smoothing import GaussianSmoothing
@@ -208,7 +207,7 @@ def _compute_max_attention_per_index(attention_maps: torch.Tensor,
         for i in indices_to_alter:
             image = attention_for_text[:, :, i]
             if smooth_attentions:
-                smoothing = GaussianSmoothing(channels=1, kernel_size=kernel_size, sigma=sigma, dim=2).half().cuda()
+                smoothing = GaussianSmoothing(channels=1, kernel_size=kernel_size, sigma=sigma, dim=2).cuda()
                 input = F.pad(image.unsqueeze(0).unsqueeze(0), (1, 1, 1, 1), mode='reflect')
                 image = smoothing(input).squeeze(0).squeeze(0)
             max_indices_list.append(image.max())
@@ -581,4 +580,4 @@ def __call__(
         if not return_dict:
             return (image, has_nsfw_concept)
 
-        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)
+        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

utils/gaussian_smoothing.py

@@ -68,7 +68,7 @@ def forward(self, input):
         Returns:
             filtered (torch.Tensor): Filtered output.
         """
-        return self.conv(input, weight=self.weight, groups=self.groups)
+        return self.conv(input, weight=self.weight.to(input.dtype), groups=self.groups)
 
 
 class AverageSmoothing(nn.Module):