fix pre-commit

detect.py

@@ -13,7 +13,7 @@
 from utils.datasets import LoadImages, LoadStreams
 from utils.general import (check_img_size, check_imshow, increment_path,
                            non_max_suppression, scale_coords, set_logging,
-                           strip_optimizer, xyxy2xywh)
+                           xyxy2xywh)
 from utils.plots import plot_one_box
 from utils.torch_utils import intersect_dicts, select_device, time_synchronized
 
@@ -98,7 +98,7 @@ def detect(save_img=False):
         if img.ndimension() == 3:
             img = img.unsqueeze(0)
 
-        # Warmup
+        # warm up
         if device.type != 'cpu' and (old_img_b != img.shape[0]
                                      or old_img_h != img.shape[2]
                                      or old_img_w != img.shape[3]):
@@ -201,7 +201,7 @@ def detect(save_img=False):
 
     if save_txt or save_img:
         s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
-        #print(f"Results saved to {save_dir}{s}")
+        # print(f"Results saved to {save_dir}{s}")
 
     print(f'Done. ({time.time() - t0:.3f}s)')
 

get_dynamic_thres.py

@@ -1,8 +1,5 @@
 import argparse
 import logging
-import os
-from pathlib import Path
-from threading import Thread
 
 import numpy as np
 import torch
@@ -29,14 +26,14 @@ def get_thres(data,
     set_logging()
     device = select_device(opt.device, batch_size=batch_size)
     if isinstance(data, str):
-        is_coco = data.endswith('coco.yaml')
+        # is_coco = data.endswith('coco.yaml')
         with open(data) as f:
             data = yaml.load(f, Loader=yaml.SafeLoader)
     check_dataset(data)  # check
     nc = int(data['nc'])  # number of classes
-    iouv = torch.linspace(0.5, 0.95,
-                          10).to(device)  # iou vector for [email protected]:0.95
-    niou = iouv.numel()
+    # iouv = torch.linspace(0.5, 0.95,
+    #                       10).to(device)  # iou vector for [email protected]:0.95
+    # niou = iouv.numel()
 
     # Load model
     model = Model(cfg, ch=3, nc=nc)  # create

hyp/hyp.finetune.dynamic.adam.yaml

@@ -2,9 +2,9 @@ lr0: 0.00001  # initial learning rate
 lrf: 1  # final OneCycleLR learning rate (lr0 * lrf)
 momentum: 0.937  # SGD momentum/Adam beta1
 weight_decay: 0.005  # optimizer weight decay 5e-4
-warmup_epochs: 0.01  # warmup epochs (fractions ok)
-warmup_momentum: 0.8  # warmup initial momentum
-warmup_bias_lr: 0.01  # warmup initial bias lr
+warmup_epochs: 0.01  # warm up epochs (fractions ok)
+warmup_momentum: 0.8  # warm up initial momentum
+warmup_bias_lr: 0.01  # warm up initial bias lr
 box: 0.05  # box loss gain
 cls: 0.3  # cls loss gain
 cls_pw: 1.0  # cls BCELoss positive_weight

hyp/hyp.scratch.p5.yaml

@@ -2,9 +2,9 @@ lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
 lrf: 0.1  # final OneCycleLR learning rate (lr0 * lrf)
 momentum: 0.937  # SGD momentum/Adam beta1
 weight_decay: 0.0005  # optimizer weight decay 5e-4
-warmup_epochs: 3.0  # warmup epochs (fractions ok)
-warmup_momentum: 0.8  # warmup initial momentum
-warmup_bias_lr: 0.1  # warmup initial bias lr
+warmup_epochs: 3.0  # warm up epochs (fractions ok)
+warmup_momentum: 0.8  # warm up initial momentum
+warmup_bias_lr: 0.1  # warm up initial bias lr
 box: 0.05  # box loss gain
 cls: 0.3  # cls loss gain
 cls_pw: 1.0  # cls BCELoss positive_weight

hyp/hyp.scratch.p6.yaml

@@ -2,9 +2,9 @@ lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
 lrf: 0.2  # final OneCycleLR learning rate (lr0 * lrf)
 momentum: 0.937  # SGD momentum/Adam beta1
 weight_decay: 0.0005  # optimizer weight decay 5e-4
-warmup_epochs: 3.0  # warmup epochs (fractions ok)
-warmup_momentum: 0.8  # warmup initial momentum
-warmup_bias_lr: 0.1  # warmup initial bias lr
+warmup_epochs: 3.0  # warm up epochs (fractions ok)
+warmup_momentum: 0.8  # warm up initial momentum
+warmup_bias_lr: 0.1  # warm up initial bias lr
 box: 0.05  # box loss gain
 cls: 0.3  # cls loss gain
 cls_pw: 1.0  # cls BCELoss positive_weight

models/common.py

@@ -6,7 +6,7 @@
 
 from utils.general import non_max_suppression
 
-##### basic ####
+# basic
 
 
 def autopad(k, p=None):  # kernel, padding
@@ -105,9 +105,7 @@ def forward(self, x):
         return x
 
 
-##### end of basic #####
-
-##### cspnet #####
+# cspnet
 
 
 class SPPCSPC(nn.Module):
@@ -132,9 +130,7 @@ def forward(self, x):
         return self.cv7(torch.cat((y1, y2), dim=1))
 
 
-##### end of cspnet #####
-
-##### yolor #####
+# yolor
 
 
 class ImplicitA(nn.Module):
@@ -165,9 +161,7 @@ def forward(self, x):
         return self.implicit * x
 
 
-##### end of yolor #####
-
-##### repvgg #####
+# repvgg
 
 
 class RepConv(nn.Module):
@@ -302,7 +296,7 @@ def fuse_conv_bn(self, conv, bn):
     def fuse_repvgg_block(self):
         if self.deploy:
             return
-        print(f'RepConv.fuse_repvgg_block')
+        print('RepConv.fuse_repvgg_block')
 
         self.rbr_dense = self.fuse_conv_bn(self.rbr_dense[0],
                                            self.rbr_dense[1])
@@ -346,9 +340,9 @@ def fuse_repvgg_block(self):
             weight_identity_expanded = torch.nn.Parameter(
                 torch.zeros_like(weight_1x1_expanded))
 
-        #print(f"self.rbr_1x1.weight = {self.rbr_1x1.weight.shape}, ")
-        #print(f"weight_1x1_expanded = {weight_1x1_expanded.shape}, ")
-        #print(f"self.rbr_dense.weight = {self.rbr_dense.weight.shape}, ")
+        # print(f"self.rbr_1x1.weight = {self.rbr_1x1.weight.shape}, ")
+        # print(f"weight_1x1_expanded = {weight_1x1_expanded.shape}, ")
+        # print(f"self.rbr_dense.weight = {self.rbr_dense.weight.shape}, ")
 
         self.rbr_dense.weight = torch.nn.Parameter(self.rbr_dense.weight +
                                                    weight_1x1_expanded +
@@ -373,9 +367,7 @@ def fuse_repvgg_block(self):
             self.rbr_dense = None
 
 
-##### end of repvgg #####
-
-##### yolov5 #####
+# yolov5
 
 
 class NMS(nn.Module):
@@ -394,9 +386,7 @@ def forward(self, x):
                                    classes=self.classes)
 
 
-##### end of yolov5 ######
-
-##### CBNet #####
+# CBNet
 
 
 class CBLinear(nn.Module):
@@ -439,9 +429,7 @@ def forward(self, xs):
         return out
 
 
-##### end of CBNet #####
-
-##### DynamicDet #####
+# DynamicDet
 
 
 def sigmoid(logits, hard=False, threshold=0.5):
@@ -484,6 +472,3 @@ def forward(self, xs, thres=0.5):
         else:
             xs = xs.sigmoid()
         return xs
-
-
-##### end of DynamicDet #####

models/yolo.py

@@ -1,29 +1,30 @@
-import argparse
 import logging
 import math
 import sys
 from copy import deepcopy
 from pathlib import Path
 
-import yaml
-
-sys.path.append('./')  # to run '$ python *.py' files in subdirectories
-logger = logging.getLogger(__name__)
 import torch
+import torch.nn as nn
+import yaml
 from torch.nn.modules.batchnorm import _BatchNorm
 
-from models.common import *
+from models.common import (NMS, SPPCSPC, AdaptiveRouter, CBFuse, CBLinear,
+                           Concat, Conv, ConvCheckpoint, ImplicitA, ImplicitM,
+                           ReOrg, RepConv, Shortcut, autoShape)
 from utils.autoanchor import check_anchor_order
-from utils.general import check_file, make_divisible, set_logging
+from utils.general import make_divisible
 from utils.torch_utils import (copy_attr, fuse_conv_and_bn, initialize_weights,
-                               model_info, scale_img, select_device,
-                               time_synchronized)
+                               model_info, scale_img, time_synchronized)
 
 try:
     import thop  # for FLOPS computation
 except ImportError:
     thop = None
 
+sys.path.append('./')  # to run '$ python *.py' files in subdirectories
+logger = logging.getLogger(__name__)
+
 
 class IDetect(nn.Module):
     stride = None  # strides computed during build
@@ -180,8 +181,10 @@ def __init__(self,
             logger.info(
                 f"Overriding model.yaml nc={self.yaml['nc']} with nc={nc}")
             self.yaml['nc'] = nc  # override yaml value
-        self.model_b, self.save_b, self.model_b2, self.save_b2, self.model_h, self.save_h, self.model_h2, self.save_h2 = parse_model(
-            deepcopy(self.yaml), ch_b=[ch])  # model, savelist
+        (self.model_b, self.save_b, self.model_b2, self.save_b2, self.model_h,
+         self.save_h, self.model_h2,
+         self.save_h2) = parse_model(deepcopy(self.yaml),
+                                     ch_b=[ch])  # model, savelist
         self.keep_input = self.yaml.get('keep_input', False)
         self.names = [str(i) for i in range(self.yaml['nc'])]  # default names
         # print([x.shape for x in self.forward(torch.zeros(1, ch, 64, 64))])
@@ -439,7 +442,7 @@ def fuse(self):  # fuse model Conv2d() + BatchNorm2d() layers
         ]:
             for m in model.modules():
                 if isinstance(m, RepConv):
-                    #print(f" fuse_repvgg_block")
+                    # print(f" fuse_repvgg_block")
                     m.fuse_repvgg_block()
                 elif type(m) is Conv and hasattr(m, 'bn'):
                     m.conv = fuse_conv_and_bn(m.conv, m.bn)  # update conv
@@ -505,7 +508,7 @@ def parse_model(d, ch_b):  # model_dict, input_channels(3)
         for j, a in enumerate(args):
             try:
                 args[j] = eval(a) if isinstance(a, str) else a  # eval strings
-            except:
+            except Exception:
                 pass
 
         n = max(round(n * gd), 1) if n > 1 else n  # depth gain
@@ -556,7 +559,7 @@ def parse_model(d, ch_b):  # model_dict, input_channels(3)
         for j, a in enumerate(args):
             try:
                 args[j] = eval(a) if isinstance(a, str) else a  # eval strings
-            except:
+            except Exception:
                 pass
 
         chs = []
@@ -633,7 +636,7 @@ def parse_model(d, ch_b):  # model_dict, input_channels(3)
         for j, a in enumerate(args):
             try:
                 args[j] = eval(a) if isinstance(a, str) else a  # eval strings
-            except:
+            except Exception:
                 pass
         chs = []
         for x in ([f] if isinstance(f, (int, str)) else f):
@@ -699,7 +702,7 @@ def parse_model(d, ch_b):  # model_dict, input_channels(3)
         for j, a in enumerate(args):
             try:
                 args[j] = eval(a) if isinstance(a, str) else a  # eval strings
-            except:
+            except Exception:
                 pass
         chs = []
         for x in ([f] if isinstance(f, (int, str)) else f):
@@ -761,5 +764,7 @@ def parse_model(d, ch_b):  # model_dict, input_channels(3)
     save_b.extend(d['b1_save'])
     save_b2.extend(d['b2_save'])
 
-    return nn.Sequential(*layers_b), sorted(save_b), nn.Sequential(*layers_b2), sorted(save_b2), \
-            nn.Sequential(*layers_h), sorted(save_h), nn.Sequential(*layers_h2), sorted(save_h2)
+    return (nn.Sequential(*layers_b),
+            sorted(save_b), nn.Sequential(*layers_b2), sorted(save_b2),
+            nn.Sequential(*layers_h), sorted(save_h),
+            nn.Sequential(*layers_h2), sorted(save_h2))

train_step1.py

@@ -26,11 +26,10 @@
 from utils.autoanchor import check_anchors
 from utils.checkpoint import get_state_dict
 from utils.datasets import create_dataloader
-from utils.general import (check_dataset, check_file, check_git_status,
-                           check_img_size, colorstr, fitness, get_latest_run,
-                           increment_path, init_seeds, labels_to_class_weights,
-                           labels_to_image_weights, one_cycle, set_logging,
-                           strip_optimizer)
+from utils.general import (check_dataset, check_file, check_img_size, colorstr,
+                           fitness, get_latest_run, increment_path, init_seeds,
+                           labels_to_class_weights, labels_to_image_weights,
+                           one_cycle, set_logging, strip_optimizer)
 from utils.loss import ComputeLoss, ComputeLossOTA, ComputeLossOTADual
 from utils.plots import plot_images, plot_lr_scheduler, plot_results
 from utils.torch_utils import (ModelEMA, intersect_dicts, is_parallel,
@@ -85,7 +84,8 @@ def train(hyp, opt, device, tb_writer=None):
         loggers['wandb'] = wandb_logger.wandb
         data_dict = wandb_logger.data_dict
         if wandb_logger.wandb:
-            weight, epochs, hyp = opt.weight, opt.epochs, opt.hyp  # WandbLogger might update weights, epochs if resuming
+            # WandbLogger might update weights, epochs if resuming
+            weight, epochs, hyp = opt.weight, opt.epochs, opt.hyp
 
     nc = 1 if opt.single_cls else int(data_dict['nc'])  # number of classes
     names = ['item'] if opt.single_cls and len(
@@ -225,8 +225,11 @@ def train(hyp, opt, device, tb_writer=None):
     # Scheduler https://arxiv.org/pdf/1812.01187.pdf
     # https://pytorch.org/docs/stable/_modules/torch/optim/lr_scheduler.html#OneCycleLR
     if opt.linear_lr:
-        lf = lambda x: (1 - x / (epochs - 1)) * (1.0 - hyp['lrf']) + hyp[
-            'lrf']  # linear
+
+        def get_linear_lr(x):
+            return (1 - x / (epochs - 1)) * (1.0 - hyp['lrf']) + hyp['lrf']
+
+        lf = get_linear_lr
     else:
         lf = one_cycle(1, hyp['lrf'], epochs)  # cosine 1->hyp['lrf']
     scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf)
@@ -362,9 +365,10 @@ def train(hyp, opt, device, tb_writer=None):
 
     # Start training
     t0 = time.time()
-    nw = max(round(hyp['warmup_epochs'] * nb),
-             1000)  # number of warmup iterations, max(3 epochs, 1k iterations)
-    # nw = min(nw, (epochs - start_epoch) / 2 * nb)  # limit warmup to < 1/2 of training
+    nw = max(
+        round(hyp['warmup_epochs'] * nb),
+        1000)  # number of warm up iterations, max(3 epochs, 1k iterations)
+    # nw = min(nw, (epochs - start_epoch) / 2 * nb)  # limit warm up to < 1/2 of training
     maps = np.zeros(nc)  # mAP per class
     results = (0, 0, 0, 0, 0, 0, 0
                )  # P, R, [email protected], [email protected], val_loss(box, obj, cls)
@@ -425,7 +429,7 @@ def train(hyp, opt, device, tb_writer=None):
             imgs = imgs.to(device, non_blocking=True).float(
             ) / 255.0  # uint8 to float32, 0-255 to 0.0-1.0
 
-            # Warmup
+            # Warm up
             if ni <= nw:
                 xi = [0, nw]  # x interp
                 # model.gr = np.interp(ni, xi, [0.0, 1.0])  # iou loss ratio (obj_loss = 1.0 or iou)
@@ -505,8 +509,8 @@ def train(hyp, opt, device, tb_writer=None):
                         ]
                     })
 
-            # end batch ------------------------------------------------------------------------------------------------
-        # end epoch ----------------------------------------------------------------------------------------------------
+            # end batch ----------------------------------------------------------------------------------------------
+        # end epoch --------------------------------------------------------------------------------------------------
 
         # Scheduler
         lr = [x['lr'] for x in optimizer.param_groups]  # for tensorboard
@@ -620,7 +624,7 @@ def train(hyp, opt, device, tb_writer=None):
                                                best_model=best_fitness == fi)
                 del ckpt
 
-        # end epoch ----------------------------------------------------------------------------------------------------
+        # end epoch --------------------------------------------------------------------------------------------------
     # end training
     if rank in [-1, 0]:
         # Plots
@@ -795,7 +799,8 @@ def train(hyp, opt, device, tb_writer=None):
                 f, Loader=yaml.SafeLoader))  # replace
         opt.cfg, opt.weight, opt.resume = os.path.relpath(
             Path(ckpt).parent.parent / 'cfg.yaml'), ckpt, True
-        opt.batch_size, opt.global_rank, opt.local_rank = opt.total_batch_size, *apriori  # reinstate
+        opt.batch_size, opt.global_rank, opt.local_rank = \
+            opt.total_batch_size, *apriori  # reinstate
         opt.save_dir = os.path.relpath(Path(ckpt).parent.parent)
         logger.info('Resuming training from %s' % ckpt)
     else: