change to inpaint

app.py

@@ -1,6 +1,6 @@
 import gradio as gr
 
-from app_haircolor import create_demo as create_demo_haircolor
+from app_haircolor_inpaint import create_demo as create_demo_haircolor
 
 with gr.Blocks(css="style.css") as demo:
     with gr.Tabs():

app_haircolor.py

@@ -38,10 +38,15 @@ pidinet_detector = pidinet_detector.to(DEVICE)
 
 canndy_detector = CannyDetector()
 
+midas_detector = MidasDetector.from_pretrained(
+    "valhalla/t2iadapter-aux-models", filename="dpt_large_384.pt", model_type="dpt_large"
+)
+midas_detector = midas_detector.to(DEVICE)
+
 adapters = MultiAdapter(
     [
         T2IAdapter.from_pretrained(
-            "TencentARC/t2iadapter_sketch_sd15v2",
+            "TencentARC/t2i-adapter-lineart-sdxl-1.0",
             torch_dtype=torch.float16,
             varient="fp16",
         ),
@@ -80,13 +85,12 @@ def image_to_image(
     run_task_time = 0
     time_cost_str = ''
     run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
-    # lineart_image = lineart_detector(input_image, int(generate_size*0.375), generate_size)
-    # run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
+    lineart_image = lineart_detector(input_image, int(generate_size*0.375), generate_size)
+    run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
     canny_image = canndy_detector(input_image, int(generate_size*0.375), generate_size)
     run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
-    sketch_image = pidinet_detector(input_image, int(generate_size*0.5), generate_size)
 
-    cond_image = [sketch_image, canny_image]
+    cond_image = [lineart_image, canny_image]
     cond_scale = [lineart_scale, canny_scale]
 
     generator = torch.Generator(device=DEVICE).manual_seed(seed)

app_haircolor_inpaint.py

@@ -9,6 +9,7 @@ from segment_utils import(
     restore_result,
 )
 from diffusers import (
+    StableDiffusionXLInpaintPipeline,
     DiffusionPipeline,
     T2IAdapter,
     MultiAdapter,
@@ -49,12 +50,12 @@ adapters = MultiAdapter(
 )
 adapters = adapters.to(torch.float16)
 
-basepipeline = DiffusionPipeline.from_pretrained(
+basepipeline = StableDiffusionXLInpaintPipeline.from_pretrained(
     BASE_MODEL,
     torch_dtype=torch.float16,
     use_safetensors=True,
-    adapter=adapters,
-    custom_pipeline="./pipelines/pipeline_sdxl_adapter_inpaint_custom.py",
+    # adapter=adapters,
+    # custom_pipeline="./pipelines/pipeline_sdxl_adapter_inpaint_custom.py",
 )
 
 basepipeline = basepipeline.to(DEVICE)
@@ -76,13 +77,13 @@ def image_to_image(
     run_task_time = 0
     time_cost_str = ''
     run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
-    lineart_image = lineart_detector(input_image, int(generate_size*0.375), generate_size)
-    run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
-    canny_image = canndy_detector(input_image, int(generate_size*0.375), generate_size)
-    run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
+    # lineart_image = lineart_detector(input_image, int(generate_size*0.375), generate_size)
+    # run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
+    # canny_image = canndy_detector(input_image, int(generate_size*0.375), generate_size)
+    # run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)
 
-    cond_image = [lineart_image, canny_image]
-    cond_scale = [lineart_scale, canny_scale]
+    # cond_image = [lineart_image, canny_image]
+    # cond_scale = [lineart_scale, canny_scale]
 
     generator = torch.Generator(device=DEVICE).manual_seed(seed)
     generated_image = basepipeline(
@@ -95,8 +96,8 @@ def image_to_image(
         width=generate_size,
         guidance_scale=guidance_scale,
         num_inference_steps=num_steps,
-        adapter_image=cond_image,
-        adapter_conditioning_scale=cond_scale,
+        # adapter_image=cond_image,
+        # adapter_conditioning_scale=cond_scale,
     ).images[0]
     
     run_task_time, time_cost_str = get_time_cost(run_task_time, time_cost_str)