Update app.py

Testing upscaler

app.py

@@ -9,7 +9,7 @@ device = 'cuda' if torch.cuda.is_available() else 'cpu'
 torch.cuda.max_memory_allocated(device=device)
 torch.cuda.empty_cache()
 
-def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, upscale, high_noise_frac):
+def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, refine, high_noise_frac, upscale):
     generator = np.random.seed(0) if seed == 0 else torch.manual_seed(seed)
        
     if Model == "PhotoReal":
@@ -17,7 +17,8 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
         pipe.enable_xformers_memory_efficient_attention()
         pipe = pipe.to(device)
         torch.cuda.empty_cache()
-        if upscale == "Yes":
+        
+        if refine == "Yes":
             refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", use_safetensors=True, torch_dtype=torch.float16, variant="fp16") if torch.cuda.is_available() else DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0")
             refiner.enable_xformers_memory_efficient_attention()
             refiner = refiner.to(device)
@@ -25,18 +26,28 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
             int_image = pipe(Prompt, negative_prompt=negative_prompt, height=height, width=width, num_inference_steps=steps, guidance_scale=scale).images
             image = refiner(Prompt, negative_prompt=negative_prompt, image=int_image, denoising_start=high_noise_frac).images[0]
             torch.cuda.empty_cache()
-            return image
-        else:
-            image = pipe(Prompt, negative_prompt=negative_prompt, height=height, width=width, num_inference_steps=steps, guidance_scale=scale).images[0]
+        
+            if upscale == "Yes":
+            upscaler = DiffusionPipeline.from_pretrained("stabilityai/sd-x2-latent-upscaler", torch_dtype=torch.float16, use_safetensors=True)
+            upscaler.enable_xformers_memory_efficient_attention()
+            upscaler = upscaler.to(device)
+            torch.cuda.empty_cache()
+            upscaled = upscaler(prompt=prompt, negative_prompt=negative_prompt, image=image, num_inference_steps=15, guidance_scale=0).images[0]
             torch.cuda.empty_cache()
+            return upscaled
+        else:
             return image
+    else:
+        image = pipe(Prompt, negative_prompt=negative_prompt, height=height, width=width, num_inference_steps=steps, guidance_scale=scale).images[0]
+        torch.cuda.empty_cache()
+        return image
     
     if Model == "Anime":
         anime = DiffusionPipeline.from_pretrained("circulus/canvers-anime-v3.8.1", torch_dtype=torch.float16, safety_checker=None) if torch.cuda.is_available() else DiffusionPipeline.from_pretrained("circulus/canvers-anime-v3.8.1")
         anime.enable_xformers_memory_efficient_attention()
         anime = anime.to(device)
         torch.cuda.empty_cache()
-        if upscale == "Yes":
+        if refine == "Yes":
             refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", use_safetensors=True, torch_dtype=torch.float16, variant="fp16") if torch.cuda.is_available() else DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0")
             refiner.enable_xformers_memory_efficient_attention()
             refiner = refiner.to(device)
@@ -55,7 +66,7 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
         disney.enable_xformers_memory_efficient_attention()
         disney = disney.to(device)
         torch.cuda.empty_cache()
-        if upscale == "Yes":
+        if refine == "Yes":
             refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", use_safetensors=True, torch_dtype=torch.float16, variant="fp16") if torch.cuda.is_available() else DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0")
             refiner.enable_xformers_memory_efficient_attention()
             refiner = refiner.to(device)
@@ -74,7 +85,7 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
         story.enable_xformers_memory_efficient_attention()
         story = story.to(device)
         torch.cuda.empty_cache()
-        if upscale == "Yes":
+        if refine == "Yes":
             refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", use_safetensors=True, torch_dtype=torch.float16, variant="fp16") if torch.cuda.is_available() else DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0")
             refiner.enable_xformers_memory_efficient_attention()
             refiner = refiner.to(device)
@@ -93,7 +104,7 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
         semi.enable_xformers_memory_efficient_attention()
         semi = semi.to(device)
         torch.cuda.empty_cache()
-        if upscale == "Yes":
+        if refine == "Yes":
             refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", use_safetensors=True, torch_dtype=torch.float16, variant="fp16") if torch.cuda.is_available() else DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0")
             refiner.enable_xformers_memory_efficient_attention()
             refiner = refiner.to(device)
@@ -112,7 +123,7 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
         animagine.enable_xformers_memory_efficient_attention()
         animagine = animagine.to(device)
         torch.cuda.empty_cache()
-        if upscale == "Yes":
+        if refine == "Yes":
             torch.cuda.empty_cache()
             torch.cuda.max_memory_allocated(device=device)
             int_image = animagine(Prompt, negative_prompt=negative_prompt, height=height, width=width, num_inference_steps=steps, guidance_scale=scale, output_type="latent").images
@@ -137,7 +148,7 @@ def genie (Model, Prompt, negative_prompt, height, width, scale, steps, seed, up
         sdxl = sdxl.to(device)   
         torch.cuda.empty_cache()
     
-        if upscale == "Yes":
+        if refine == "Yes":
             torch.cuda.max_memory_allocated(device=device)
             torch.cuda.empty_cache()
             image = sdxl(Prompt, negative_prompt=negative_prompt, height=height, width=width, num_inference_steps=steps, guidance_scale=scale, output_type="latent").images
@@ -164,7 +175,8 @@ gr.Interface(fn=genie, inputs=[gr.Radio(['PhotoReal', 'Anime', 'Disney', 'StoryB
                                gr.Slider(25, maximum=100, value=50, step=25, label='Number of Iterations'), 
                                gr.Slider(minimum=0, step=1, maximum=9999999999999999, randomize=True, label='Seed: 0 is Random'), 
                                gr.Radio(["Yes", "No"], label='SDXL 1.0 Refiner: Use if the Image has too much Noise', value='No'),
-                               gr.Slider(minimum=.9, maximum=.99, value=.95, step=.01, label='Refiner Denoise Start %')],
+                               gr.Slider(minimum=.9, maximum=.99, value=.95, step=.01, label='Refiner Denoise Start %'),
+                               gr.Radio(["Yes", "No"], label = 'SD 2.0 X2 Latent Upscaler?', value="No")],
              outputs=gr.Image(label='Generated Image'), 
              title="Manju Dream Booth V1.6 with SDXL 1.0 Refiner - GPU", 
              description="<br><br><b/>Warning: This Demo is capable of producing NSFW content.",