Upload app_i2i.py

app_i2i.py

@@ -0,0 +1,119 @@
+import os
+import numpy as np
+import gradio as gr
+from utils.i2i import i2i_gen
+
+MAX_SEED = np.iinfo(np.int32).max
+MIN_IMAGE_SIZE = int(os.getenv("MIN_IMAGE_SIZE", "512"))
+MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "2048"))
+
+with gr.Blocks(
+    title="🪄 LayerDiffuse - Flux version (Image to Image)",
+    theme="CultriX/gradio-theme"
+) as demo:
+    gr.Markdown(
+        """
+        # 🪄 LayerDiffuse - Flux version (Image to Image)
+
+        A Flux version implementation of LayerDiffuse for image-to-image generation.
+        
+        Upload an image with transparency (PNG with alpha channel) and transform it with a text prompt.
+        """
+    )
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            input_image = gr.Image(
+                label="Input Image",
+                type="pil",
+                tool="editor",
+                sources=["upload", "webcam", "clipboard"],
+                image_mode="RGBA",
+                height=512
+            )
+            
+            prompt = gr.Text(
+                label="Prompt",
+                info="Describe what you want to generate",
+                placeholder="E.g: a beautiful landscape with mountains, high quality"
+            )
+            
+            with gr.Row():
+                seed = gr.Slider(
+                    label="Seed",
+                    minimum=0,
+                    maximum=MAX_SEED,
+                    step=1,
+                    value=0,
+                    randomize=True
+                )
+                
+                strength = gr.Slider(
+                    label="Strength",
+                    minimum=0.0,
+                    maximum=1.0,
+                    step=0.01,
+                    value=0.8,
+                    info="How much to transform the image (0 = no change, 1 = complete change)"
+                )
+            
+            with gr.Row():
+                guidance_scale = gr.Slider(
+                    label="Guidance scale",
+                    minimum=1,
+                    maximum=20,
+                    step=0.1,
+                    value=7.0,
+                )
+                
+                num_inference_steps = gr.Slider(
+                    label="Steps",
+                    minimum=10,
+                    maximum=100,
+                    step=1,
+                    value=50,
+                )
+            
+            generate_btn = gr.Button("Generate", variant="primary")
+        
+        with gr.Column(scale=1):
+            output_image = gr.Image(
+                label="Result",
+                type="pil",
+                height=512
+            )
+    
+    # Set up the generation flow
+    gr.on(
+        triggers=[generate_btn.click],
+        fn=lambda: gr.update(interactive=False, value="Generating..."),
+        outputs=generate_btn,
+        api_name=False
+    ).then(
+        fn=i2i_gen,
+        inputs=[
+            input_image,
+            prompt,
+            seed,
+            guidance_scale,
+            num_inference_steps,
+            strength
+        ],
+        outputs=output_image
+    ).then(
+        fn=lambda: gr.update(interactive=True, value="Generate"),
+        outputs=generate_btn,
+        api_name=False
+    )
+    
+    # Add examples
+    gr.Examples(
+        examples=[
+            ["./imgs/causal_cut.png", "a beautiful landscape with mountains, high quality"],
+            ["./imgs/causal_cut.png", "a futuristic cityscape, cyberpunk style, detailed"],
+        ],
+        inputs=[input_image, prompt],
+    )
+
+if __name__ == "__main__":
+    demo.queue(max_size=20).launch(show_error=True)