directamente

app.py

@@ -7,6 +7,7 @@ from share_btn import community_icon_html, loading_icon_html, share_js
 from diffuserslocal.src.diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_ldm3d_inpaint import StableDiffusionLDM3DInpaintPipeline
 from PIL import Image
 import numpy as np
+import cv2
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
@@ -35,7 +36,6 @@ else:
 
 def estimate_depth(image):
 
-    image = np.array(image)
     input_batch = transform(image).to(device)
 
     with torch.no_grad():
@@ -68,17 +68,17 @@ def predict(dict, depth, prompt="", negative_prompt="", guidance_scale=7.5, step
         negative_prompt = None
     scheduler_class_name = scheduler.split("-")[0]
 
-    init_image = dict["image"].convert("RGB").resize((512, 512))
+    init_image = cv2.resize(dict["image"], (512, 512))
     
     if (depth == None):
         depth_image = estimate_depth(init_image)
     else:
-        depth_image = depth.convert("L")
+        depth_image = depth
     
     scheduler = getattr(diffusers, scheduler_class_name)
     pipe.scheduler = scheduler.from_pretrained("Intel/ldm3d-4c", subfolder="scheduler")
     
-    mask = dict["mask"].convert("RGB").resize((512, 512))
+    mask = cv2.resize(dict["mask"], (512, 512))
     depth_image = depth_image.resize((512, 512))
     
     output = pipe(prompt = prompt, negative_prompt=negative_prompt, image=init_image, mask_image=mask, depth_image=depth_image, guidance_scale=guidance_scale, num_inference_steps=int(steps), strength=strength)
@@ -128,10 +128,8 @@ with image_blocks as demo:
     gr.HTML(read_content("header.html"))
     with gr.Row():
                 with gr.Column():
-                    image = gr.Image(source='upload', tool='sketch', elem_id="image_upload", type="pil", label="Upload",height=400)
-                    depth = gr.Image(source='upload', elem_id="depth_upload", type="pil", label="Upload",height=400, image_mode="L")
-                    
-                    print(depth)
+                    image = gr.Image(source='upload', tool='sketch', elem_id="image_upload", type="numpy", label="Upload",height=400)
+                    depth = gr.Image(source='upload', elem_id="depth_upload", type="numpy", label="Upload",height=400)
 
                     with gr.Row(elem_id="prompt-container", mobile_collapse=False, equal_height=True):
                         with gr.Row():