inital deployment1

app.py

@@ -3,11 +3,12 @@ import cv2
 import torch
 import numpy as np
 from diffusers import StableDiffusionImg2ImgPipeline
+from transformers import AutoProcessor, AutoModel
 from PIL import Image
 
-# Load the Stable Diffusion Model
+# Load the Real-Time Latent Consistency Model
 device = "cuda" if torch.cuda.is_available() else "cpu"
-pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5").to(device)
+realtime_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("radames/Real-Time-Latent-Consistency-Model").to(device)
 
 def apply_color_filter(frame, filter_type="None"):
     """Apply a color filter to the frame."""
@@ -23,14 +24,14 @@ def apply_color_filter(frame, filter_type="None"):
     return frame
 
 def process_frame(frame, filter_type="None", prompt="A futuristic landscape"):
-    """Process a single frame by applying a color filter and diffusion model."""
+    """Process a single frame by applying a color filter and real-time latent consistency model."""
     frame = apply_color_filter(frame, filter_type)
     
     # Convert frame to PIL image
     image = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)).resize((512, 512))
     
-    # Apply Stable Diffusion model
-    result = pipe(prompt=prompt, image=image, strength=0.5, guidance_scale=7.5).images[0]
+    # Apply Real-Time Latent Consistency Model
+    result = realtime_pipe(prompt=prompt, image=image, strength=0.5, guidance_scale=7.5).images[0]
     return np.array(result)
 
 def video_stream(filter_type, prompt):
@@ -45,21 +46,17 @@ def video_stream(filter_type, prompt):
         yield frame  # Return processed frame
     cap.release()
 
+# Create Gradio App
+with gr.Blocks() as demo:
+    gr.Markdown("## 🎨 Real-Time AI-Enhanced Webcam using Latent Consistency Model")
+    
+    with gr.Row():
+        webcam_feed = gr.Camera(streaming=True, label="Live Webcam")
+        processed_image = gr.Image(label="AI-Enhanced Webcam Feed")
+    
+    filter_selector = gr.Radio(["None", "Red", "Green", "Blue"], label="Color Filter")
+    prompt_input = gr.Textbox(label="Real-Time Latent Consistency Model Prompt", value="A futuristic landscape")
+    
+    webcam_feed.change(fn=video_stream, inputs=[filter_selector, prompt_input], outputs=processed_image)
 
-
-
-
-iface = gr.Interface(
-    fn=video_stream,
-    inputs=[
-        gr.Video( streaming=True),  # Use browser webcam
-        gr.Radio(["None", "Red", "Green", "Blue"], label="Color Filter"),
-        gr.Textbox(label="Stable Diffusion Prompt", value="A futuristic landscape"),
-    ],
-    outputs=gr.Image(label="AI-Enhanced Webcam Feed"),
-    live=True,
-)
-
-
-if __name__ == "__main__":
-    iface.launch(share=True)
+demo.launch(share=True)