Update app.py

app.py

@@ -10,10 +10,11 @@ import gradio as gr
 import numpy as np
 import spaces
 import torch
+import cv2
 from PIL import Image
 from io import BytesIO
 from diffusers.utils import load_image
-from diffusers import AutoencoderKL, DiffusionPipeline, AutoPipelineForImage2Image, AutoPipelineForInpainting
+from diffusers import StableDiffusionControlNetPipeline, ControlNetModel, AutoencoderKL, DiffusionPipeline, AutoPipelineForImage2Image, AutoPipelineForInpainting
 
 DESCRIPTION = "# Run any LoRA or SD Model"
 if not torch.cuda.is_available():
@@ -27,6 +28,7 @@ ENABLE_USE_LORA = os.getenv("ENABLE_USE_LORA", "1") == "1"
 ENABLE_USE_VAE = os.getenv("ENABLE_USE_VAE", "1") == "1"
 ENABLE_USE_IMG2IMG = os.getenv("ENABLE_USE_IMG2IMG", "1") == "1"
 ENABLE_USE_INPAINTING = os.getenv("ENABLE_USE_INPAINTING", "1") == "1"
+ENABLE_USE_CONTROLNET = os.getenv("ENABLE_USE_CONTROLNET", "1") == "1"
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
@@ -59,6 +61,7 @@ def generate(
     lora_scale: float = 0.7,
     use_img2img: bool = False,
     use_inpainting: bool = False,
+    use_controlnet: bool = False,
     url = '',
     img_url = '',
     mask_url = '',
@@ -84,6 +87,19 @@ def generate(
 
             image_init = load_image(img_url)
             mask_image = load_image(mask_url)
+
+        if use_controlnet:
+            controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny", torch_dtype=torch.float16)
+            pipe = StableDiffusionControlNetPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", controlnet=controlnet, torch_dtype=torch.float16)
+
+            image = load_image(img_url)
+            image = np.array(image)
+            
+            # get canny image
+            image = cv2.Canny(image, 100, 200)
+            image = image[:, :, None]
+            image = np.concatenate([image, image, image], axis=2)
+            canny_image = Image.fromarray(image)
             
         if use_lora:
             pipe.load_lora_weights(lora)
@@ -107,6 +123,21 @@ def generate(
     if not use_negative_prompt_2:
         negative_prompt_2 = None  # type: ignore
 
+    if use_controlnet:
+        image = pipe(
+            prompt=prompt,
+            image=image,
+            control_image=canny_image,
+            negative_prompt=negative_prompt,
+            prompt_2=prompt_2,
+            width=width,
+            height=height,
+            negative_prompt_2=negative_prompt_2,
+            guidance_scale=guidance_scale_base,
+            num_inference_steps=num_inference_steps_base,
+            generator=generator,
+        ).images[0]
+        return image
     if use_inpainting:
         image = pipe(
             prompt=prompt,
@@ -185,6 +216,7 @@ with gr.Blocks(theme=gr.themes.Soft(), css="style.css") as demo:
         result = gr.Image(label="Result", show_label=False)
     with gr.Accordion("Advanced options", open=False):
         with gr.Row():
+            use_controlnet = gr.Checkbox(label='Use Controlnet'), value=False, visible=ENABLE_USE_CONTROLNET)
             use_inpainting = gr.Checkbox(label='Use Inpainting', value=False, visible=ENABLE_USE_INPAINTING)
             use_img2img = gr.Checkbox(label='Use Img2Img', value=False, visible=ENABLE_USE_IMG2IMG)
             use_vae = gr.Checkbox(label='Use VAE', value=False, visible=ENABLE_USE_VAE)
@@ -319,6 +351,13 @@ with gr.Blocks(theme=gr.themes.Soft(), css="style.css") as demo:
         queue=False,
         api_name=False,
     )
+    use_controlnet.change(
+        fn=lambda x: gr.update(visible=x),
+        inputs=use_controlnet,
+        outputs=img_url,
+        queue=False,
+        api_name=False,
+    )
 
     gr.on(
         triggers=[
@@ -360,6 +399,7 @@ with gr.Blocks(theme=gr.themes.Soft(), css="style.css") as demo:
             url,
             img_url,
             mask_url,
+            use_controlnet,
         ],
         outputs=result,
         api_name="run",