Update app.py

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
 from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, AutoencoderKL
-from diffusers import DDIMScheduler, EulerAncestralDiscreteScheduler
 import torch
 import numpy as np
 import cv2
@@ -12,11 +11,9 @@ device = "cuda"
 precision = torch.float16
 
 
-eulera_scheduler = EulerAncestralDiscreteScheduler.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", subfolder="scheduler")
-
 # 🏗️ Load ControlNet model for Canny edge detection
 controlnet = ControlNetModel.from_pretrained(
-    "thepowerfuldeez/sd21-controlnet-canny",
+    "diffusers/controlnet-canny-sdxl-1.0",
     torch_dtype=precision
 )
 
@@ -28,22 +25,22 @@ pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
     controlnet=controlnet,
     vae=vae,
     torch_dtype=precision,
-    scheduler=eulera_scheduler,
 )
 
+pipe.enable_model_cpu_offload()
 pipe.to(device)
 
-# ✨ Enable attention slicing for CPU optimization
-#pipe.enable_attention_slicing()
+## to move in the function
+controlnet_conditioning_scale = 0.5  # recommended for good generalization
 
 # 📸 Edge detection function using OpenCV (Canny)
 @spaces.GPU
 def apply_canny(image, low_threshold, high_threshold):
-    image = np.array(image.convert("RGB"))
-    gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
-    edges = cv2.Canny(gray, low_threshold, high_threshold)
-    edges = np.stack([edges] * 3, axis=-1)  # Convert to RGB format
-    return Image.fromarray(edges)
+    image = np.array(image)
+    image = cv2.Canny(image, low_threshold, high_threshold)
+    image = image[:, :, None]
+    image = np.concatenate([image, image, image], axis=2)
+    return Image.fromarray(image)
 
 # 🎨 Image generation function
 @spaces.GPU
@@ -57,6 +54,7 @@ def generate_image(prompt, input_image, low_threshold, high_threshold, strength,
         image=edge_detected,
         num_inference_steps=30,
         guidance_scale=guidance,
+        controlnet_conditioning_scale=controlnet_conditioning_scale,
         strength=strength
     ).images[0]