Merge Ultralytics Code Refactor

Automated merge of Ultralytics Code Refactor

README.md

@@ -93,7 +93,7 @@ If you use this repository or the associated tools and datasets in your research
 
 # 👥 Contribute
 
-🤝 We love contributions from the community! Our open-source projects thrive on your help. To start contributing, please check out our [Contributing Guide](https://docs.ultralytics.com/help/contributing). Additionally, we'd love to hear from you through our [Survey](https://ultralytics.com/survey?utm_source=github&utm_medium=social&utm_campaign=Survey). It's a way to **impact** the future of our projects. A big shoutout and thank you 🙏 to all our contributors!
+🤝 We love contributions from the community! Our open-source projects thrive on your help. To start contributing, please check out our [Contributing Guide](https://docs.ultralytics.com/help/contributing). Additionally, we'd love to hear from you through our [Survey](https://www.ultralytics.com/survey?utm_source=github&utm_medium=social&utm_campaign=Survey). It's a way to **impact** the future of our projects. A big shoutout and thank you 🙏 to all our contributors!
 
 <!-- Image with SVG format can be troublesome in some markdown viewers -->
 
@@ -104,12 +104,12 @@ If you use this repository or the associated tools and datasets in your research
 
 At Ultralytics, we provide two different licensing options to suit various use cases:
 
-- **AGPL-3.0 License**: The [AGPL-3.0 License](https://www.gnu.org/licenses/agpl-3.0.html) is an [OSI-approved](https://opensource.org/licenses/) open-source format that's best suited for students, researchers, and enthusiasts to promote collaboration and knowledge sharing. The full terms can be found in the [LICENSE](https://github.com/ultralytics/ultralytics/blob/main/LICENSE) file.
-- **Enterprise License**: If you're looking for a commercial application of our software and models, the Enterprise License enables integration into commercial products while bypassing the open-source stipulations of the AGPL-3.0. For embedding our solutions into your commercial offerings, please contact us through [Ultralytics Licensing](https://ultralytics.com/license).
+- **AGPL-3.0 License**: The [AGPL-3.0 License](https://www.gnu.org/licenses/agpl-3.0.html) is an [OSI-approved](https://opensource.org/license) open-source format that's best suited for students, researchers, and enthusiasts to promote collaboration and knowledge sharing. The full terms can be found in the [LICENSE](https://github.com/ultralytics/ultralytics/blob/main/LICENSE) file.
+- **Enterprise License**: If you're looking for a commercial application of our software and models, the Enterprise License enables integration into commercial products while bypassing the open-source stipulations of the AGPL-3.0. For embedding our solutions into your commercial offerings, please contact us through [Ultralytics Licensing](https://www.ultralytics.com/license).
 
 # 📬 Contact
 
-🐞 For reporting bugs or suggesting new features, please open an issue on our [GitHub Issues](https://github.com/ultralytics/xview-yolov3/issues) page. And if you have questions or fancy engaging with us, join our vibrant [Discord](https://ultralytics.com/discord) community!
+🐞 For reporting bugs or suggesting new features, please open an issue on our [GitHub Issues](https://github.com/ultralytics/xview-yolov3/issues) page. And if you have questions or fancy engaging with us, join our vibrant [Discord](https://discord.com/invite/ultralytics) community!
 
 <br>
 <div align="center">

detect.py

@@ -100,10 +100,10 @@ def detect(opt):
         ni = int(math.ceil(img.shape[1] / length))  # up-down
         nj = int(math.ceil(img.shape[2] / length))  # left-right
         for i in range(ni):  # for i in range(ni - 1):
-            print("row %g/%g: " % (i, ni), end="")
+            print(f"row {i:g}/{ni:g}: ", end="")
 
             for j in range(nj):  # for j in range(nj if i==0 else nj - 1):
-                print("%g " % j, end="", flush=True)
+                print(f"{j:g} ", end="", flush=True)
 
                 # forward scan
                 y2 = min((i + 1) * length, img.shape[1])
@@ -166,7 +166,7 @@ def detect(opt):
 
     # Iterate through images and save plot of detections
     for img_i, (path, detections) in enumerate(zip(imgs, img_detections)):
-        print("image %g: '%s'" % (img_i, path))
+        print(f"image {img_i:g}: '{path}'")
 
         if opt.plot_flag:
             img = cv2.imread(path)
@@ -192,7 +192,7 @@ def detect(opt):
             with open(results_path.replace(".bmp", ".tif") + ".txt", "a") as file:
                 for i in unique_classes:
                     n = (detections[:, -1].cpu() == i).sum()
-                    print("%g %ss" % (n, classes[int(i)]))
+                    print(f"{n:g} {classes[int(i)]}s")
 
                 for x1, y1, x2, y2, conf, cls_conf, cls_pred in detections:
                     # Rescale coordinates to original dimensions
@@ -207,11 +207,11 @@ def detect(opt):
                     # write to file
                     xvc = xview_indices2classes(int(cls_pred))  # xview class
                     # if (xvc != 21) & (xvc != 72):
-                    file.write(("%g %g %g %g %g %g \n") % (x1, y1, x2, y2, xvc, cls_conf * conf))
+                    file.write(f"{x1:g} {y1:g} {x2:g} {y2:g} {xvc:g} {cls_conf * conf:g} \n")
 
                     if opt.plot_flag:
                         # Add the bbox to the plot
-                        label = "%s %.2f" % (classes[int(cls_pred)], cls_conf) if cls_conf > 0.05 else None
+                        label = f"{classes[int(cls_pred)]} {cls_conf:.2f}" if cls_conf > 0.05 else None
                         color = bbox_colors[int(np.where(unique_classes == int(cls_pred))[0])]
                         plot_one_box([x1, y1, x2, y2], img, label=label, color=color, line_thickness=1)
 
@@ -236,7 +236,7 @@ def __init__(self, num_classes=60):
         """Initializes a ConvNetb model with configurable number of classes, defaulting to 60, and a series of
         convolutional layers.
         """
-        super(ConvNetb, self).__init__()
+        super().__init__()
         n = 64  # initial convolution size
         self.layer1 = nn.Sequential(
             nn.Conv2d(3, n, kernel_size=3, stride=1, padding=1, bias=False), nn.BatchNorm2d(n), nn.LeakyReLU()

models.py

@@ -73,15 +73,15 @@ class EmptyLayer(nn.Module):
 
     def __init__(self):
         """Initializes a placeholder layer for 'route' and 'shortcut' in YOLO architecture."""
-        super(EmptyLayer, self).__init__()
+        super().__init__()
 
 
 class YOLOLayer(nn.Module):
     """Processes YOLO detection layers with specified anchors, classes, and image dimensions for object detection."""
 
     def __init__(self, anchors, nC, img_dim, anchor_idxs):
         """Initializes YOLO layer with given anchors, number of classes, image dimensions, and anchor indexes."""
-        super(YOLOLayer, self).__init__()
+        super().__init__()
 
         anchors = list(anchors)
         nA = len(anchors)
@@ -232,7 +232,7 @@ def __init__(self, config_path, img_size=416):
         """Initializes Darknet model with a configuration path and optional image size, parsing and creating model
         modules.
         """
-        super(Darknet, self).__init__()
+        super().__init__()
         self.module_defs = parse_model_config(config_path)
         self.module_defs[0]["height"] = img_size
         self.hyperparams, self.module_list = create_modules(self.module_defs)
@@ -294,7 +294,7 @@ def forward(self, x, targets=None, requestPrecision=False, weight=None, epoch=No
 
 def parse_model_config(path):
     """Parses the yolo-v3 layer configuration file and returns module definitions."""
-    file = open(path, "r")
+    file = open(path)
     lines = file.read().split("\n")
     lines = [x for x in lines if x and not x.startswith("#")]
     lines = [x.rstrip().lstrip() for x in lines]  # get rid of fringe whitespaces

scoring/matching.py

@@ -20,7 +20,7 @@
 from scoring.rectangle import Rectangle
 
 
-class Matching(object):
+class Matching:
     """Matching class."""
 
     def __init__(self, groundtruth_rects, rects):

scoring/rectangle.py

@@ -14,7 +14,7 @@
 """
 
 
-class Rectangle(object):
+class Rectangle:
     """Rectangle class."""
 
     def __init__(self, xmin, ymin, xmax, ymax):

scoring/score.py

@@ -66,7 +66,7 @@ def get_labels(fname):
 
     coords = np.zeros((len(data["features"]), 4))
     chips = np.zeros((len(data["features"])), dtype="object")
-    classes = np.zeros((len(data["features"])))
+    classes = np.zeros(len(data["features"]))
 
     for i in tqdm(range(len(data["features"]))):
         if data["features"][i]["properties"]["bounds_imcoords"] != []:
@@ -169,7 +169,7 @@ def score(path_predictions, path_groundtruth, path_output, iou_threshold=0.5):
         fname = file.split(".txt")[0]
         pchips.append(fname)
 
-        with open(path_predictions + file, "r") as f:
+        with open(path_predictions + file) as f:
             arr = np.array(list(csv.reader(f, delimiter=" ")))
             if arr.shape[0] == 0:
                 # If the file is empty, we fill it in with an array of zeros
@@ -202,13 +202,13 @@ def score(path_predictions, path_groundtruth, path_output, iou_threshold=0.5):
     max_gt_cls = 100
 
     if set(pchips).issubset(set(gt_unique)):
-        raise ValueError("The prediction files {%s} are not in the ground truth." % str(set(pchips) - (set(gt_unique))))
+        raise ValueError(f"The prediction files {{{str(set(pchips) - (set(gt_unique)))}}} are not in the ground truth.")
 
     print("Number of Predictions: %d" % num_preds)
     print("Number of GT: %d" % np.sum(gt_classes.shape))
 
     per_file_class_data = {i: [[], []] for i in gt_unique}
-    num_gt_per_cls = np.zeros((max_gt_cls))
+    num_gt_per_cls = np.zeros(max_gt_cls)
 
     attempted = np.zeros(100)
     for file_ind in range(len(pchips)):
@@ -456,15 +456,17 @@ def score(path_predictions, path_groundtruth, path_output, iou_threshold=0.5):
     vals["f1"] = 2 / ((1 / (np.spacing(1) + vals["map_score"])) + (1 / (np.spacing(1) + vals["mar_score"])))
 
     print(
-        "mAP: %f | mAP score: %f | mAR: %f | F1: %f" % (vals["map"], vals["map_score"], vals["mar_score"], vals["f1"])
+        "mAP: {:f} | mAP score: {:f} | mAR: {:f} | F1: {:f}".format(
+            vals["map"], vals["map_score"], vals["mar_score"], vals["f1"]
+        )
     )
 
     # with open(path_output + '/score.txt', 'w') as f:
     #     f.write(str("%.8f" % vals['map']))
     #
     with open(f"{path_output}/metrics.txt", "w") as f:
         for key, value in vals.items():
-            f.write("%s %f\n" % (str(key), value))
+            f.write(f"{str(key)} {value:f}\n")
         # for key in vals.keys():
         #     f.write("%f\n" % (vals[key]))
         for i in range(len(v2)):

train.py

@@ -153,8 +153,8 @@ def main(opt):
                 k = (metrics[0] + metrics[2]) > 0
                 mean_recall = recall[k].mean() if k.sum() > 0 else 0
                 s = ("%10s%10s" + "%10.3g" * 14) % (
-                    "%g/%g" % (epoch, opt.epochs - 1),
-                    "%g/%g" % (i, len(dataloader) - 1),
+                    f"{epoch:g}/{opt.epochs - 1:g}",
+                    f"{i:g}/{len(dataloader) - 1:g}",
                     rloss["x"],
                     rloss["y"],
                     rloss["w"],
@@ -214,7 +214,7 @@ def main(opt):
 
     # Save final model
     dt = time.time() - t0
-    print("Finished %g epochs in %.2fs (%.2fs/epoch)" % (epoch, dt, dt / (epoch + 1)))
+    print(f"Finished {epoch:g} epochs in {dt:.2f}s ({dt / (epoch + 1):.2f}s/epoch)")
 
 
 if __name__ == "__main__":

utils/datasets.py

@@ -128,7 +128,7 @@ def __next__(self):
         labels_all = []
         for files_index in range(ia, ib):
             # img_path = self.files[self.shuffled_vector[files_index]]  # BGR
-            img_path = "%s/%g.tif" % (self.path, self.shuffled_vector[files_index])
+            img_path = f"{self.path}/{self.shuffled_vector[files_index]:g}.tif"
             # img_path = '/Users/glennjocher/Downloads/DATA/xview/train_images/2294.bmp'
 
             img0 = cv2.imread(img_path)
@@ -383,7 +383,7 @@ def convert_tif2bmp(p="/Users/glennjocher/Downloads/DATA/xview/val_images_bmp"):
 
     files = sorted(glob.glob(f"{p}/*.tif"))
     for i, f in enumerate(files):
-        print("%g/%g" % (i + 1, len(files)))
+        print(f"{i + 1:g}/{len(files):g}")
 
         img = cv2.imread(f)
 

utils/utils.py

@@ -12,7 +12,7 @@
 
 def load_classes(path):
     """Loads class labels at 'path'."""
-    fp = open(path, "r")
+    fp = open(path)
     return fp.read().split("\n")[:-1]
 
 
@@ -26,7 +26,7 @@ def modelinfo(model):
         print(
             "%4g %70s %9s %12g %20s %12g %12g" % (i, name, p.requires_grad, p.numel(), list(p.shape), p.mean(), p.std())
         )
-    print("\n%g layers, %g parameters, %g gradients" % (i + 1, nparams, ngradients))
+    print(f"\n{i + 1:g} layers, {nparams:g} parameters, {ngradients:g} gradients")
 
 
 def xview_classes2indices(classes):  # remap xview classes 11-94 to 0-61
@@ -698,9 +698,9 @@ def secondary_class_detection(x, y, w, h, img, model, device):
     with torch.no_grad():
         classes = []
         nB = n // 1000 + 1
-        print("%g batches..." % nB, end="")
+        print(f"{nB:g} batches...", end="")
         for i in range(nB):
-            print("%g " % i, end="")
+            print(f"{i:g} ", end="")
             j0 = int(i * 1000)
             j1 = int(min(j0 + 1000, n))
             im = images[j0:j1]
@@ -729,9 +729,9 @@ def createChips():
         print(counter)
 
         if platform == "darwin":  # macos
-            img = cv2.imread("/Users/glennjocher/Downloads/DATA/xview/train_images/%g.tif" % i)
+            img = cv2.imread(f"/Users/glennjocher/Downloads/DATA/xview/train_images/{i:g}.tif")
         else:  # gcp
-            img = cv2.imread("../train_images/%g.tif" % i)
+            img = cv2.imread(f"../train_images/{i:g}.tif")
 
         for j in np.nonzero(mat["id"] == i)[0]:
             c, x1, y1, x2, y2 = mat["targets"][j]

utils/utils_xview.py

@@ -6,7 +6,7 @@
 
 def xview_class2name(classes):
     """Converts numerical class IDs to their corresponding names using 'data/xview.names'."""
-    with open("data/xview.names", "r") as f:
+    with open("data/xview.names") as f:
         x = f.readlines()
     return x[classes].replace("\n", "")