Update app.py

app.py

@@ -129,7 +129,8 @@ def preprocess_image(image: Image.Image,
                     style_name: str = "",
                     num_steps: int = 25,
                     guidance_scale: float = 5,
-                    controlnet_conditioning_scale: float = 1.0,) -> Image.Image:
+                    controlnet_conditioning_scale: float = 1.0,
+                    do_preprocess: bool = True) -> Image.Image:
     """
     Preprocess the input image.
 
@@ -140,32 +141,35 @@ def preprocess_image(image: Image.Image,
         Image.Image: The preprocessed image.
     """
         
-    width, height  = image['composite'].size
-    ratio = np.sqrt(1024. * 1024. / (width * height))
-    new_width, new_height = int(width * ratio), int(height * ratio)
-    image = image['composite'].resize((new_width, new_height))
-
-    print("image:",type(image))
-
-    prompt, negative_prompt = apply_style(style_name, prompt, negative_prompt)
+    if do_preprocess:
+        width, height  = image['composite'].size
+        ratio = np.sqrt(1024. * 1024. / (width * height))
+        new_width, new_height = int(width * ratio), int(height * ratio)
+        image = image['composite'].resize((new_width, new_height))
     
-    print("params:", prompt, negative_prompt, style_name, num_steps, guidance_scale, controlnet_conditioning_scale)
-    image = pipe_control(
-        prompt=prompt,
-        negative_prompt=negative_prompt,
-        image=image,
-        num_inference_steps=num_steps,
-        controlnet_conditioning_scale=controlnet_conditioning_scale,
-        guidance_scale=guidance_scale,
-        width=new_width,
-        height=new_height).images[0]
+        print("image:",type(image))
     
+        prompt, negative_prompt = apply_style(style_name, prompt, negative_prompt)
         
-    processed_image = pipeline.preprocess_image(image)
-    return processed_image
+        print("params:", prompt, negative_prompt, style_name, num_steps, guidance_scale, controlnet_conditioning_scale)
+        image = pipe_control(
+            prompt=prompt,
+            negative_prompt=negative_prompt,
+            image=image,
+            num_inference_steps=num_steps,
+            controlnet_conditioning_scale=controlnet_conditioning_scale,
+            guidance_scale=guidance_scale,
+            width=new_width,
+            height=new_height).images[0]
+        
+            
+        processed_image = pipeline.preprocess_image(image)
+        return processed_image, False
+    else:
+        return image, False
 
 
-def preprocess_images(images: List[Tuple[Image.Image, str]]) -> List[Image.Image]:
+def preprocess_images(images: List[Tuple[Image.Image, str]], do_preprocess = True) -> List[Image.Image]:
     """
     Preprocess a list of input images.
     
@@ -177,7 +181,7 @@ def preprocess_images(images: List[Tuple[Image.Image, str]]) -> List[Image.Image
     """
     images = [image[0] for image in images]
     processed_images = [pipeline.preprocess_image(image) for image in images]
-    return processed_images
+    return processed_images, False
 
 
 def pack_state(gs: Gaussian, mesh: MeshExtractResult) -> dict:
@@ -327,6 +331,9 @@ def extract_glb(
     return glb_path, glb_path
 
 
+def reset_do_preprocess():
+    return True
+    
 @spaces.GPU
 def extract_gaussian(state: dict, req: gr.Request) -> Tuple[str, str]:
     """
@@ -377,11 +384,8 @@ def split_image(image: Image.Image) -> List[Image.Image]:
 
 with gr.Blocks(delete_cache=(600, 600), js=js_func) as demo:
     gr.Markdown("""
-    ## Image to 3D Asset with [TRELLIS](https://trellis3d.github.io/)
-    * Upload an image and click "Generate" to create a 3D asset. If the image has alpha channel, it be used as the mask. Otherwise, we use `rembg` to remove the background.
-    * If you find the generated 3D asset satisfactory, click "Extract GLB" to extract the GLB file and download it.
-    
-    ✨New: 1) Experimental multi-image support. 2) Gaussian file extraction.
+    ## Sketch to 3D Asset with [TRELLIS](https://trellis3d.github.io/)
+    * draw or upload a sketch and click "Generate" to create a 3D asset. 
     """)
     
     with gr.Row():
@@ -438,6 +442,7 @@ with gr.Blocks(delete_cache=(600, 600), js=js_func) as demo:
                 download_gs = gr.DownloadButton(label="Download Gaussian", interactive=False)  
     
     is_multiimage = gr.State(False)
+    do_preprocess = gr.State(True)
     output_buf = gr.State()
 
     #Example images at the bottom of the page
@@ -476,15 +481,19 @@ with gr.Blocks(delete_cache=(600, 600), js=js_func) as demo:
         outputs=[is_multiimage, single_image_example, multiimage_example]
     )
     
-    image_prompt.upload(
-        preprocess_image,
-        inputs=[image_prompt, prompt],
-        outputs=[image_prompt],
+    # image_prompt.upload(
+    #     preprocess_image,
+    #     inputs=[image_prompt, prompt, negative_prompt, style, do_preprocess],
+    #     outputs=[image_prompt, do_preprocess],
+    # )
+    image_prompt.change(
+        reset_do_preprocess,
+        outputs=[do_preprocess]
     )
     multiimage_prompt.upload(
         preprocess_images,
         inputs=[multiimage_prompt],
-        outputs=[multiimage_prompt],
+        outputs=[multiimage_prompt, do_preprocess],
     )
 
     generate_btn.click(
@@ -493,12 +502,12 @@ with gr.Blocks(delete_cache=(600, 600), js=js_func) as demo:
         outputs=[seed],
     ).then(
         preprocess_image,
-        inputs=[image_prompt, prompt, negative_prompt, style],
-        outputs=[image_prompt],
+        inputs=[image_prompt, prompt, negative_prompt, style, do_preprocess],
+        outputs=[image_prompt, do_preprocess],
     ).then(
         image_to_3d,
-        inputs=[image_prompt, multiimage_prompt, is_multiimage, seed, ss_guidance_strength, ss_sampling_steps, slat_guidance_strength, slat_sampling_steps, multiimage_algo],
-        outputs=[output_buf, video_output],
+        inputs=[image_prompt, multiimage_prompt, is_multiimage, seed, ss_guidance_strength, ss_sampling_steps, slat_guidance_strength, slat_sampling_steps, multiimage_algo, do_preprocess],
+        outputs=[output_buf, video_output, do_preprocess],
     ).then(
         lambda: tuple([gr.Button(interactive=True), gr.Button(interactive=True)]),
         outputs=[extract_glb_btn, extract_gs_btn],
@@ -552,7 +561,6 @@ if __name__ == "__main__":
         controlnet=controlnet,
         vae=vae,
         torch_dtype=torch.float16,
-        # scheduler=eulera_scheduler,
     )
     pipe_control.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe_control.scheduler.config)
     pipe_control.to(device)