Explicit Imports (#498)

* expand imports

* optimize

* miss

* fix

utils/datasets.py

@@ -14,7 +14,7 @@ from PIL import Image, ExifTags
 from torch.utils.data import Dataset
 from tqdm import tqdm
 
-from utils.utils import xyxy2xywh, xywh2xyxy, torch_distributed_zero_first
+from utils.general import xyxy2xywh, xywh2xyxy, torch_distributed_zero_first
 
 help_url = 'https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data'
 img_formats = ['.bmp', '.jpg', '.jpeg', '.png', '.tif', '.tiff', '.dng']

utils/general.py

@@ -18,10 +18,11 @@ import torch
 import torch.nn as nn
 import torchvision
 import yaml
+from scipy.cluster.vq import kmeans
 from scipy.signal import butter, filtfilt
 from tqdm import tqdm
 
-from . import torch_utils  # torch_utils, google_utils
+from utils.torch_utils import init_seeds, is_parallel
 
 # Set printoptions
 torch.set_printoptions(linewidth=320, precision=5, profile='long')
@@ -47,7 +48,7 @@ def torch_distributed_zero_first(local_rank: int):
 def init_seeds(seed=0):
     random.seed(seed)
     np.random.seed(seed)
-    torch_utils.init_seeds(seed=seed)
+    init_seeds(seed=seed)
 
 
 def get_latest_run(search_dir='./runs'):
@@ -505,7 +506,7 @@ def compute_loss(p, targets, model):  # predictions, targets, model
 
 def build_targets(p, targets, model):
     # Build targets for compute_loss(), input targets(image,class,x,y,w,h)
-    det = model.module.model[-1] if torch_utils.is_parallel(model) else model.model[-1]  # Detect() module
+    det = model.module.model[-1] if is_parallel(model) else model.model[-1]  # Detect() module
     na, nt = det.na, targets.shape[0]  # number of anchors, targets
     tcls, tbox, indices, anch = [], [], [], []
     gain = torch.ones(7, device=targets.device)  # normalized to gridspace gain
@@ -779,7 +780,6 @@ def kmean_anchors(path='./data/coco128.yaml', n=9, img_size=640, thr=4.0, gen=10
     wh = wh0[(wh0 >= 2.0).any(1)]  # filter > 2 pixels
 
     # Kmeans calculation
-    from scipy.cluster.vq import kmeans
     print('Running kmeans for %g anchors on %g points...' % (n, len(wh)))
     s = wh.std(0)  # sigmas for whitening
     k, dist = kmeans(wh / s, n, iter=30)  # points, mean distance