a little milestone

README.md

@@ -76,7 +76,12 @@ When run `SUBMISSION=/path/to/csv/file.csv make release-csv`, If you encounter t
 ## TODO
 - [x] finish classify + segment model
 - [x] finish create_submission.py 
-- [ ] finish demo.py
-- [ ] finish loss.py
-- [ ] finish choose_threshold
-- [ ] finish data enhancement
+- [x] finish demo.py
+- [x] finish loss.py
+- [x] finish choose_threshold
+- [x] finish data augmentation
+- [ ] EfficientB4( w/ ASPP)
+- [ ] code review(validation dice, threshold dice)
+- [ ] choose fold
+- [ ] ensemble
+- [ ] early stopping automaticly

config.py

@@ -19,8 +19,8 @@ def get_seg_config():
         hwp: 6 MXQ: 12
         '''
         # parser.add_argument('--image_size', type=int, default=768, help='image size')
-        parser.add_argument('--batch_size', type=int, default=12, help='batch size')
-        parser.add_argument('--epoch', type=int, default=50, help='epoch')
+        parser.add_argument('--batch_size', type=int, default=24, help='batch size')
+        parser.add_argument('--epoch', type=int, default=65, help='epoch')
 
         parser.add_argument('--augmentation_flag', type=bool, default=True, help='if true, use augmentation method in train set')
         parser.add_argument('--n_splits', type=int, default=5, help='n_splits_fold')
@@ -30,14 +30,14 @@ def get_seg_config():
         parser.add_argument('--width', type=int, default=None, help='the width of cropped image')
 
         # model set 
-        parser.add_argument('--model_name', type=str, default='unet_se_resnext50_32x4d', \
+        parser.add_argument('--model_name', type=str, default='unet_resnet34', \
             help='unet_resnet34/unet_se_resnext50_32x4d')
 
         # model hyper-parameters
         parser.add_argument('--class_num', type=int, default=4)
         parser.add_argument('--resume', type=str, default=0, help='Resuming from specified weight')
         parser.add_argument('--num_workers', type=int, default=8)
-        parser.add_argument('--lr', type=float, default=4e-5, help='init lr')
+        parser.add_argument('--lr', type=float, default=5e-5, help='init lr')
         parser.add_argument('--weight_decay', type=float, default=0, help='weight_decay in optimizer')
 
         # dataset 

demo.py

@@ -36,11 +36,17 @@ def demo(n_splits, use_segment_only, model_name, mean, std, show_truemask_flag,
 
     # start prediction
     if show_truemask_flag:
-        for images, masks in tqdm(dataloader):
+        for samples in tqdm(dataloader):
+            if len(samples) == 0:
+                continue
+            images, masks = samples[0], samples[1]
             results = model(images).detach().cpu().numpy()
             pred_show(images, results, mean, std, targets=masks, flag=show_truemask_flag, auto_flag=auto_flag)
     else:
-        for fnames, images in tqdm(dataloader):
+        for fnames, samples in tqdm(dataloader):
+            if len(samples) == 0:
+                continue
+            images, masks = samples[0], samples[1]
             results = model(images).detach().cpu().numpy()
             pred_show(images, results, mean, std, targets=None, flag=show_truemask_flag, auto_flag=auto_flag)
 

train_classify.py

@@ -83,13 +83,14 @@ def train(self, train_loader, valid_loader):
 
                 # 保存到tensorboard，每一步存储一个
                 self.writer.add_scalar('train_loss', loss.item(), global_step+i)
-
-                descript = "Fold: %d, Train Loss: %.7f, lr: %.7f" % (self.fold, loss.item(), self.lr)
+                params_groups_lr = str()
+                for group_ind, param_group in enumerate(optimizer.param_groups):
+                    params_groups_lr = params_groups_lr + 'params_group_%d' % (group_ind) + ': %.12f, ' % (param_group['lr'])
+                descript = "Fold: %d, Train Loss: %.7f, lr: %s" % (self.fold, loss.item(), params_groups_lr)
                 tbar.set_description(desc=descript)
 
             # 每一个epoch完毕之后，执行学习率衰减
             lr_scheduler.step()
-            self.lr = lr_scheduler.get_lr()
             global_step += len(train_loader)
 
             # Print the log info

train_segment.py

@@ -47,8 +47,8 @@ def __init__(self, config, fold):
         self.solver = Solver(self.model)
 
         # 加载损失函数
-        self.criterion = torch.nn.BCEWithLogitsLoss()
-        # self.criterion = MultiClassesSoftBCEDiceLoss(classes_num=4, size_average=True, weight=[0.75, 0.25])
+        # self.criterion = torch.nn.BCEWithLogitsLoss()
+        self.criterion = MultiClassesSoftBCEDiceLoss(classes_num=4, size_average=True, weight=[1.0, 1.0])
 
         # 保存json文件和初始化tensorboard
         TIMESTAMP = "{0:%Y-%m-%dT%H-%M-%S-%d}".format(datetime.datetime.now(), fold)
@@ -94,13 +94,14 @@ def train(self, train_loader, valid_loader):
 
                 # 保存到tensorboard，每一步存储一个
                 self.writer.add_scalar('train_loss', loss.item(), global_step+i)
-
-                descript = "Fold: %d, Train Loss: %.7f, lr: %.7f" % (self.fold, loss.item(), self.lr)
+                params_groups_lr = str()
+                for group_ind, param_group in enumerate(optimizer.param_groups):
+                    params_groups_lr = params_groups_lr + 'params_group_%d' % (group_ind) + ': %.12f, ' % (param_group['lr'])
+                descript = "Fold: %d, Train Loss: %.7f, lr: %s" % (self.fold, loss.item(), params_groups_lr)
                 tbar.set_description(desc=descript)
 
             # 每一个epoch完毕之后，执行学习率衰减
             lr_scheduler.step()
-            self.lr = lr_scheduler.get_lr()
             global_step += len(train_loader)
 
             # Print the log info

utils/loss.py

@@ -123,7 +123,7 @@ def __init__(self, size_average=True, weight=[0.2, 0.8]):
     def forward(self, input, target):
         soft_bce_loss = self.bce_loss(input, target)
         soft_dice_loss = self.softdiceloss(input, target)
-        loss = 5.0 * soft_bce_loss + soft_dice_loss
+        loss = 0.7 * soft_bce_loss + 0.3 * soft_dice_loss
 
         return loss