Update app.py

app.py

@@ -152,17 +152,12 @@ def model_process(image, mask):
         cv2_flag='INPAINT_NS',
         cv2_radius=5,
     )
-    '''
-    config = ldm_steps=25 ldm_sampler='plms' zits_wireframe=True hd_strategy='Original' hd_strategy_crop_margin=196 
-    hd_strategy_crop_trigger_size=1280              hd_strategy_resize_limit=2048
-    '''
-    
-    print(f'config = {config}')
+   
+    # print(f'config = {config}')
 
-    # if config.sd_seed == -1:
-    #    config.sd_seed = random.randint(1, 999999999)
+    if config.sd_seed == -1:
+        config.sd_seed = random.randint(1, 999999999)
     
-    config.sd_seed = 42
     # logger.info(f"Origin image shape: {original_shape}")
     print(f"Origin image shape: {original_shape}")
     image = resize_max_size(image, size_limit=size_limit, interpolation=interpolation)
@@ -177,10 +172,12 @@ def model_process(image, mask):
         
     start = time.time()
     res_np_img = model(image, mask, config)
-    logger.info(f"process time: {(time.time() - start) * 1000}ms")
+    logger.info(f"process time: {(time.time() - start) * 1000}ms, {res_np_img.shape}")
 
     torch.cuda.empty_cache()
 
+    return Image.fromarray(res_np_img)
+    '''
     if alpha_channel is not None:
         if alpha_channel.shape[:2] != res_np_img.shape[:2]:
             alpha_channel = cv2.resize(
@@ -192,8 +189,8 @@ def model_process(image, mask):
 
     ext = get_image_ext(origin_image_bytes)
     return ext
+    '''
     
-'''
 model = ModelManager(
         name='lama',
         device=device,
@@ -203,7 +200,6 @@ model = ModelManager(
         # sd_run_local=True,
         # callback=diffuser_callback,
     )
-'''
 
 '''
 pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-inpainting", dtype=torch.float16, revision="fp16", use_auth_token=auth_token).to(device)
@@ -235,8 +231,8 @@ def predict(dict):
     '''
     image = Image.fromarray(dict["image"])
     mask = Image.fromarray(dict["mask"])
-    output1 = model_process(dict["image"], dict["mask"])
-    output = mask #output.images[0]
+    output = model_process(dict["image"], dict["mask"])
+    # output = mask #output.images[0]
     # output = pipe(prompt = prompt, image=init_image, mask_image=mask,guidance_scale=7.5)
     
     return output #, gr.update(visible=True), gr.update(visible=True), gr.update(visible=True)