Ultralytics Code Refactor https://ultralytics.com/actions

README.md

@@ -1,5 +1,5 @@
 <br>
-<img src="https://raw.githubusercontent.com/ultralytics/assets/main/logo/Ultralytics_Logotype_Original.svg" width="320">
+<a href="https://ultralytics.com" target="_blank"><img src="https://raw.githubusercontent.com/ultralytics/assets/main/logo/Ultralytics_Logotype_Original.svg" width="320" alt="Ultralytics logo"></a>
 
 # Introduction :wave:
 
@@ -104,7 +104,7 @@ For bug reports, feature requests, and contributions, head to [GitHub Issues](ht
   <img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="3%" alt="space">
   <a href="https://www.tiktok.com/@ultralytics"><img src="https://github.com/ultralytics/assets/raw/main/social/logo-social-tiktok.png" width="3%" alt="Ultralytics TikTok"></a>
   <img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="3%" alt="space">
-  <a href="https://www.instagram.com/ultralytics/"><img src="https://github.com/ultralytics/assets/raw/main/social/logo-social-instagram.png" width="3%" alt="Ultralytics Instagram"></a>
+  <a href="https://ultralytics.com/bilibili"><img src="https://github.com/ultralytics/assets/raw/main/social/logo-social-bilibili.png" width="3%" alt="Ultralytics Instagram"></a>
   <img src="https://github.com/ultralytics/assets/raw/main/social/logo-transparent.png" width="3%" alt="space">
   <a href="https://ultralytics.com/discord"><img src="https://github.com/ultralytics/assets/raw/main/social/logo-social-discord.png" width="3%" alt="Ultralytics Discord"></a>
 </div>

detect.py

@@ -27,6 +27,7 @@
 
 
 def detect(opt):
+    """Perform object detection on images using the specified model and configurations."""
     if opt.plot_flag:
         os.system(f"rm -rf {opt.output_folder}_img")
         os.makedirs(f"{opt.output_folder}_img", exist_ok=True)
@@ -215,6 +216,7 @@ def detect(opt):
 
 class ConvNetb(nn.Module):
     def __init__(self, num_classes=60):
+        """Initializes the ConvNetb class with a specified number of output classes."""
         super(ConvNetb, self).__init__()
         n = 64  # initial convolution size
         self.layer1 = nn.Sequential(
@@ -247,6 +249,7 @@ def __init__(self, num_classes=60):
         self.fully_convolutional = nn.Conv2d(n * 16, 60, kernel_size=4, stride=1, padding=0, bias=True)
 
     def forward(self, x):  # 500 x 1 x 64 x 64
+        """Performs the forward pass by sequentially applying model layers to input tensor x."""
         x = self.layer1(x)
         x = self.layer2(x)
         x = self.layer3(x)

models.py

@@ -72,13 +72,15 @@ class EmptyLayer(nn.Module):
     """Placeholder for 'route' and 'shortcut' layers."""
 
     def __init__(self):
+        """Initializes an EmptyLayer module."""
         super(EmptyLayer, self).__init__()
 
 
 class YOLOLayer(nn.Module):
     # YOLO Layer 0
 
     def __init__(self, anchors, nC, img_dim, anchor_idxs):
+        """Initializes a YOLOLayer module with anchors, class count, image dimensions, and anchor indices."""
         super(YOLOLayer, self).__init__()
 
         anchors = list(anchors)
@@ -107,6 +109,9 @@ def __init__(self, anchors, nC, img_dim, anchor_idxs):
 
     # @profile
     def forward(self, p, targets=None, requestPrecision=False, weight=None, epoch=None):
+        """Performs a forward pass in the model with given input tensors, target data, precision flag, weights, and
+        epoch info.
+        """
         FT = torch.cuda.FloatTensor if p.is_cuda else torch.FloatTensor
         torch.device("cuda:0" if p.is_cuda else "cpu")
         # weight = xview_class_weights(range(60)).to(device)
@@ -225,6 +230,7 @@ class Darknet(nn.Module):
     """YOLOv3 object detection model."""
 
     def __init__(self, config_path, img_size=416):
+        """Initialize the YOLOv3 model with configuration path and optional image size (default 416)."""
         super(Darknet, self).__init__()
         self.module_defs = parse_model_config(config_path)
         self.module_defs[0]["height"] = img_size
@@ -233,6 +239,9 @@ def __init__(self, config_path, img_size=416):
         self.loss_names = ["loss", "x", "y", "w", "h", "conf", "cls", "nGT", "TP", "FP", "FPe", "FN", "TC"]
 
     def forward(self, x, targets=None, requestPrecision=False, weight=None, epoch=None):
+        """Perform a forward pass through the network, optionally computing loss and returning outputs, targets, and
+        other metrics.
+        """
         is_training = targets is not None
         output = []
         self.losses = defaultdict(float)

utils/datasets.py

@@ -14,6 +14,9 @@
 
 class ImageFolder:  # for eval-only
     def __init__(self, path, batch_size=1, img_size=416):
+        """Initialize ImageFolder with a path, batch size, and image size, setting up file paths for image data
+        loading.
+        """
         if os.path.isdir(path):
             self.files = sorted(glob.glob(f"{path}/*.*"))
         elif os.path.isfile(path):
@@ -30,10 +33,12 @@ def __init__(self, path, batch_size=1, img_size=416):
         self.rgb_std = np.array([29.99, 24.498, 22.046], dtype=np.float32).reshape((3, 1, 1))
 
     def __iter__(self):
+        """Initialize and return the iterable object with a reset count."""
         self.count = -1
         return self
 
     def __next__(self):
+        """Advance to the next item in the iterable sequence, updating the counter and returning the next image path."""
         self.count += 1
         if self.count == self.nB:
             raise StopIteration
@@ -51,11 +56,13 @@ def __next__(self):
         return [img_path], img
 
     def __len__(self):
+        """Return the number of batches 'nB'."""
         return self.nB  # number of batches
 
 
 class ListDataset:  # for training
     def __init__(self, path, batch_size=1, img_size=608, targets_path=""):
+        """Initialize ListDataset with image path, batch size, image size, and targets path for training."""
         self.path = path
         self.files = sorted(glob.glob(f"{path}/*.bmp"))
         self.nF = len(self.files)  # number of image files
@@ -88,6 +95,7 @@ def __init__(self, path, batch_size=1, img_size=608, targets_path=""):
         # self.rgb_std = np.array([69.095, 66.369, 64.236], dtype=np.float32).reshape((1, 3, 1, 1))
 
     def __iter__(self):
+        """Initialize shuffle vector and reset count for iterating over dataset."""
         self.count = -1
         # self.shuffled_vector = np.random.permutation(self.nF)  # shuffled vector
         self.shuffled_vector = np.random.choice(
@@ -97,6 +105,9 @@ def __iter__(self):
 
     # @profile
     def __next__(self):
+        """Return the next item in the sequence, incrementing the count and stopping iteration if the count equals
+        nB.
+        """
         self.count += 1
         if self.count == self.nB:
             raise StopIteration
@@ -268,10 +279,12 @@ def __next__(self):
         return torch.from_numpy(img_all), labels_all
 
     def __len__(self):
+        """Returns the total number of batches (self.nB)."""
         return self.nB  # number of batches
 
 
 def xview_classes2indices(classes):  # remap xview classes 11-94 to 0-61
+    """Remaps xView classes (11-94) to indices (0-61) for the given list of classes."""
     indices = [
         -1,
         -1,
@@ -373,6 +386,7 @@ def xview_classes2indices(classes):  # remap xview classes 11-94 to 0-61
 
 
 def resize_square(img, height=416, color=(0, 0, 0)):  # resizes a rectangular image to a padded square
+    """Resizes a rectangular image to a padded square with a specified height and color."""
     shape = img.shape[:2]  # shape = [height, width]
     ratio = float(height) / max(shape)
     new_shape = [round(shape[0] * ratio), round(shape[1] * ratio)]
@@ -455,6 +469,7 @@ def random_affine(
 
 
 def convert_tif2bmp(p="/Users/glennjocher/Downloads/DATA/xview/val_images_bmp"):
+    """Convert all .tif files to .bmp format in the specified directory path."""
     import glob
 
     import cv2

utils/utils.py

@@ -17,6 +17,9 @@ def load_classes(path):
 
 
 def modelinfo(model):
+    """Displays model layer details including name, gradient status, number of parameters, shape, mean, and standard
+    deviation.
+    """
     nparams = sum(x.numel() for x in model.parameters())
     ngradients = sum(x.numel() for x in model.parameters() if x.requires_grad)
     print("\n%4s %70s %9s %12s %20s %12s %12s" % ("", "name", "gradient", "parameters", "shape", "mu", "sigma"))
@@ -29,12 +32,14 @@ def modelinfo(model):
 
 
 def xview_class2name(classes):
+    """Converts an xView class index to its corresponding name by reading from 'data/xview.names' file."""
     with open("data/xview.names", "r") as f:
         x = f.readlines()
     return x[classes].replace("\n", "")
 
 
 def xview_indices2classes(indices):  # remap xview classes 11-94 to 0-61
+    """Remaps xView class indices from 11-94 to 0-61."""
     class_list = [
         11,
         12,
@@ -101,6 +106,7 @@ def xview_indices2classes(indices):  # remap xview classes 11-94 to 0-61
 
 
 def xview_class_weights(indices):  # weights of each class in the training set, normalized to mu = 1
+    """Returns normalized class weights for given indices in the xView dataset using torch.FloatTensor."""
     weights = 1 / torch.FloatTensor(
         [
             74,
@@ -170,6 +176,7 @@ def xview_class_weights(indices):  # weights of each class in the training set,
 
 
 def xview_feedback_weights(indices):
+    """Calculate normalization weights for given xView feedback indices using pre-defined values."""
     weights = 1 / torch.FloatTensor(
         [
             0,
@@ -240,6 +247,7 @@ def xview_feedback_weights(indices):
 
 
 def plot_one_box(x, im, color=None, label=None, line_thickness=None):
+    """Draws a bounding box with optional label on an image using specified coordinates, color, and line thickness."""
     tl = line_thickness or round(0.003 * max(im.shape[:2]))
     color = color or [random.randint(0, 255) for _ in range(3)]
     c1, c2 = (int(x[0]), int(x[1])), (int(x[2]), int(x[3]))
@@ -253,6 +261,7 @@ def plot_one_box(x, im, color=None, label=None, line_thickness=None):
 
 
 def weights_init_normal(m):
+    """Initialize the weights of convolutional and batch normalization layers with a normal distribution."""
     classname = m.__class__.__name__
     if classname.find("Conv") != -1:
         torch.nn.init.normal_(m.weight.data, 0.0, 0.03)
@@ -262,6 +271,7 @@ def weights_init_normal(m):
 
 
 def xyxy2xywh(box):
+    """Convert bounding box format from (x1, y1, x2, y2) to (x_center, y_center, width, height)."""
     xywh = np.zeros(box.shape)
     xywh[:, 0] = (box[:, 0] + box[:, 2]) / 2
     xywh[:, 1] = (box[:, 1] + box[:, 3]) / 2
@@ -574,7 +584,9 @@ def non_max_suppression(prediction, conf_thres=0.5, nms_thres=0.4, mat=None, img
 
 # @profile
 def secondary_class_detection(x, y, w, h, img, model, device):
-    # 1. create 48-pixel squares from each chip
+    """Detect secondary classes from input image chips using a specified model and device, returning class
+    predictions.
+    """
     img = np.ascontiguousarray(img.transpose([1, 2, 0]))  # torch to cv2
     height = 64
 
@@ -626,6 +638,7 @@ def secondary_class_detection(x, y, w, h, img, model, device):
 
 
 def createChips():
+    """Generates image chips from unique images and saves them with labels into an HDF5 file."""
     from sys import platform
 
     import cv2
@@ -684,6 +697,7 @@ def createChips():
 
 
 def plotResults():
+    """Plot results from multiple text files for various metrics using matplotlib."""
     import matplotlib.pyplot as plt
     import numpy as np