2

app.py

@@ -118,7 +118,7 @@ def inference(
         print("Error: input_image is empty.")
         return None
     
-    mask_image = None
+    
 
     # input_image, width, height = scale_image(input_image, image_scale)
     # control_mode='HED'
@@ -139,19 +139,25 @@ def inference(
     control_image, width, height = rescale_image(input_image, image_scale, 16)
     control_image = control_image.resize((1024, 1024))
 
-    print("DEBUG: processor running")
+    print("DEBUG: control_image_torch")
+    sample_size = [height, width]
     control_image = processor(control_image, to_pil=True)
     control_image = control_image.resize((width, height))
-
-    sample_size = [height, width]
-    if mask_image is not None:
-        mask_image = get_image_latent(mask_image, sample_size=sample_size)[:, :1, 0]
-    else:
-        mask_image = torch.ones([1, 1, sample_size[0], sample_size[1]]) * 255
-
-    print("DEBUG: control_image_torch")
     control_image_torch = get_image_latent(control_image, sample_size=sample_size)[:, :, 0]
 
+    # mask_image = None
+    # inpaint_image = None
+    
+    # if mask_image is not None:
+    #     mask_image = get_image_latent(mask_image, sample_size=sample_size)[:, :1, 0]
+    # else:
+    #     mask_image = torch.ones([1, 1, sample_size[0], sample_size[1]]) * 255
+    
+    # if inpaint_image is not None:
+    #     inpaint_image = get_image_latent(inpaint_image, sample_size=sample_size)[:, :, 0]
+    # else:
+    #     inpaint_image = torch.zeros([1, 3, sample_size[0], sample_size[1]])
+
     # generation
     if randomize_seed: seed = random.randint(0, MAX_SEED)
     generator = torch.Generator().manual_seed(seed)
@@ -163,6 +169,8 @@ def inference(
         width=width,
         generator=generator,
         guidance_scale=guidance_scale,
+        image               = None,
+        mask_image          = None,
         control_image=control_image_torch,
         num_inference_steps=num_inference_steps,
         control_context_scale=control_context_scale,