Update code following deepsource checks.

src/ml_utils.py

@@ -25,14 +25,16 @@ transforms = A.Compose(
 )
 
 
-def cells_to_bboxes(predictions: torch.tensor, anchors: torch.tensor, s: int, is_preds: bool = True) -> torch.tensor:
+def cells_to_bboxes(
+    predictions: torch.tensor, tensor_anchors: torch.tensor, s: int, is_preds: bool = True
+) -> torch.tensor:
     """
     Scale the predictions coming from the model_files to
     be relative to the entire image such that they for example later
     can be plotted or.
     Args:
         predictions: tensor of size (N, 3, S, S, num_classes+5)
-        anchors: the anchors used for the predictions
+        tensor_anchors: the anchors used for the predictions
         s: the number of cells the image is divided in on the width (and height)
         is_preds: whether the input is predictions or the true bounding boxes
     Returns:
@@ -40,12 +42,12 @@ def cells_to_bboxes(predictions: torch.tensor, anchors: torch.tensor, s: int, is
                       object score, bounding box coordinates
     """
     batch_size = predictions.shape[0]
-    num_anchors = len(anchors)
+    num_anchors = len(tensor_anchors)
     box_predictions = predictions[..., 1:5]
     if is_preds:
-        anchors = anchors.reshape(1, len(anchors), 1, 1, 2)
+        tensor_anchors = tensor_anchors.reshape(1, len(tensor_anchors), 1, 1, 2)
         box_predictions[..., 0:2] = torch.sigmoid(box_predictions[..., 0:2])
-        box_predictions[..., 2:] = torch.exp(box_predictions[..., 2:]) * anchors
+        box_predictions[..., 2:] = torch.exp(box_predictions[..., 2:]) * tensor_anchors
         scores = torch.sigmoid(predictions[..., 0:1])
         best_class = torch.argmax(predictions[..., 5:], dim=-1).unsqueeze(-1)
     else:
@@ -79,8 +81,6 @@ def non_max_suppression(
         list: bboxes after performing NMS given a specific IoU threshold
     """
 
-    assert type(bboxes) == list
-
     bboxes = [box for box in bboxes if box[1] > threshold]
     bboxes = sorted(bboxes, key=lambda x: x[1], reverse=True)
     bboxes_after_nms = []
@@ -177,6 +177,7 @@ def predict(
     # postprocess. In fact, we could remove indexing with idx here, as there is a single image.
     # But I prefer to keep it so that this code could be easier changed for cases with batch size > 1
     bboxes: List[List] = [[] for _ in range(1)]
+    idx = 0
     for i in range(3):
         S = logits[i].shape[2]
         # it could be better to initialize anchors inside the function, but I don't want to do it for every prediction.

src/model_architecture.py

@@ -60,8 +60,6 @@ class Net(nn.Module):
         super().__init__()
         self.num_classes = 12
         self.in_channels = 3
-        # self.config = cfg.model_files.params.config
-        # self.config = [i if i[0] != '(' else literal_eval(i) for i in self.config]
         self.config = [
             (32, 3, 1),
             (64, 3, 2),
@@ -88,7 +86,6 @@ class Net(nn.Module):
             (256, 3, 1),
             'S',
         ]
-        # print('self.config', self.config)
         self.layers = self._create_conv_layers()
 
     def forward(self, x):
@@ -98,7 +95,6 @@ class Net(nn.Module):
             if isinstance(layer, ScalePrediction):
                 outputs.append(layer(x))
                 continue
-            # print(layer, x.shape)
             x = layer(x)
 
             if isinstance(layer, ResidualBlock) and layer.num_repeats == 8:
@@ -115,7 +111,6 @@ class Net(nn.Module):
         in_channels = self.in_channels
 
         for module in self.config:
-            # print(module, type(module))
             if isinstance(module, tuple):
                 out_channels, kernel_size, stride = module
                 layers.append(

src/utils.py

@@ -28,7 +28,7 @@ def plot_img_with_rects(
     ax.imshow(img)
 
     # Create a Rectangle patch
-    for _, rect in enumerate([b for b in boxes if b[1] > threshold]):
+    for _, rect in enumerate(b for b in boxes if b[1] > threshold):
         label, _, xc, yc, w, h = rect
         xc, yc, w, h = xc * coef, yc * coef, w * coef, h * coef
         # the coordinates from center-based to left top corner

st_app.py

@@ -2,7 +2,6 @@ import logging
 
 import numpy as np
 import streamlit as st
-import tomli as tomllib
 from PIL import Image
 from streamlit_drawable_canvas import st_canvas