diff --git a/rembg/bg.py b/rembg/bg.py
index 6354a03..3911177 100644
--- a/rembg/bg.py
+++ b/rembg/bg.py
@@ -8,7 +8,7 @@ from pymatting.foreground.estimate_foreground_ml import estimate_foreground_ml
 from pymatting.util.util import stack_images
 from scipy.ndimage.morphology import binary_erosion
 
-from .u2net import detect
+from .detect import load_model, predict
 
 
 def alpha_matting_cutout(
@@ -71,11 +71,11 @@ def naive_cutout(img, mask):
 @functools.lru_cache(maxsize=None)
 def get_model(model_name):
     if model_name == "u2netp":
-        return detect.load_model(model_name="u2netp")
+        return load_model(model_name="u2netp")
     if model_name == "u2net_human_seg":
-        return detect.load_model(model_name="u2net_human_seg")
+        return load_model(model_name="u2net_human_seg")
     else:
-        return detect.load_model(model_name="u2net")
+        return load_model(model_name="u2net")
 
 
 def resize_image(img, width, height):
@@ -105,7 +105,7 @@ def remove(
         img = resize_image(img, width, height)
 
     model = get_model(model_name)
-    mask = detect.predict(model, np.array(img)).convert("L")
+    mask = predict(model, np.array(img)).convert("L")
 
     if alpha_matting:
         try:
diff --git a/rembg/u2net/data_loader.py b/rembg/data_loader.py
similarity index 100%
rename from rembg/u2net/data_loader.py
rename to rembg/data_loader.py
diff --git a/rembg/u2net/detect.py b/rembg/detect.py
similarity index 95%
rename from rembg/u2net/detect.py
rename to rembg/detect.py
index 62c442d..9b73c8d 100644
--- a/rembg/u2net/detect.py
+++ b/rembg/detect.py
@@ -15,8 +15,8 @@ from skimage import transform
 from torchvision import transforms
 from tqdm import tqdm
 
-from . import data_loader, u2net
-
+from .data_loader import RescaleT, ToTensorLab
+from .u2net import U2NETP, U2NET
 
 def download_file_from_google_drive(id, fname, destination):
     head, tail = os.path.split(destination)
@@ -55,7 +55,7 @@ def load_model(model_name: str = "u2net"):
     hashfile = lambda f: md5(open(f, "rb").read()).hexdigest()
 
     if model_name == "u2netp":
-        net = u2net.U2NETP(3, 1)
+        net = U2NETP(3, 1)
         path = os.environ.get(
             "U2NETP_PATH",
             os.path.expanduser(os.path.join("~", ".u2net", model_name + ".pth")),
@@ -71,7 +71,7 @@ def load_model(model_name: str = "u2net"):
             )
 
     elif model_name == "u2net":
-        net = u2net.U2NET(3, 1)
+        net = U2NET(3, 1)
         path = os.environ.get(
             "U2NET_PATH",
             os.path.expanduser(os.path.join("~", ".u2net", model_name + ".pth")),
@@ -82,12 +82,12 @@ def load_model(model_name: str = "u2net"):
         ):
             download_file_from_google_drive(
                 "1ao1ovG1Qtx4b7EoskHXmi2E9rp5CHLcZ",
-                "u2net.pth",
+                "pth",
                 path,
             )
 
     elif model_name == "u2net_human_seg":
-        net = u2net.U2NET(3, 1)
+        net = U2NET(3, 1)
         path = os.environ.get(
             "U2NET_PATH",
             os.path.expanduser(os.path.join("~", ".u2net", model_name + ".pth")),
@@ -149,7 +149,7 @@ def preprocess(image):
         label = label[:, :, np.newaxis]
 
     transform = transforms.Compose(
-        [data_loader.RescaleT(320), data_loader.ToTensorLab(flag=0)]
+        [RescaleT(320), ToTensorLab(flag=0)]
     )
     sample = transform({"imidx": np.array([0]), "image": image, "label": label})
 
diff --git a/rembg/u2net/u2net.py b/rembg/u2net.py
similarity index 100%
rename from rembg/u2net/u2net.py
rename to rembg/u2net.py
diff --git a/rembg/u2net/__init__.py b/rembg/u2net/__init__.py
deleted file mode 100644
index e69de29..0000000