changes

cocdownload.py

@@ -0,0 +1,13 @@
+import fiftyone as fo
+import fiftyone.zoo as foz
+
+dataset = foz.load_zoo_dataset(
+    "coco/coco-2014",
+    splits=["validation", "test"],
+    label_types=["segmentations"],
+    max_samples=1000,
+    shuffle=True,
+    format=".jpg"
+)
+session = fo.launch_app(dataset)
+session.dataset = dataset

coco_splits/split_-1.json

@@ -0,0 +1 @@
+{"val": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80]}

coco_splits/split_0.json

@@ -0,0 +1 @@
+{"val": [1, 5, 9, 13, 17, 21, 25, 29, 33, 37, 41, 45, 49, 53, 57, 61, 65, 69, 73, 77], "train": [2, 3, 4, 6, 7, 8, 10, 11, 12, 14, 15, 16, 18, 19, 20, 22, 23, 24, 26, 27, 28, 30, 31, 32, 34, 35, 36, 38, 39, 40, 42, 43, 44, 46, 47, 48, 50, 51, 52, 54, 55, 56, 58, 59, 60, 62, 63, 64, 66, 67, 68, 70, 71, 72, 74, 75, 76, 78, 79, 80]}

coco_splits/split_1.json

@@ -0,0 +1 @@
+{"val": [2, 6, 10, 14, 18, 22, 26, 30, 34, 38, 42, 46, 50, 54, 58, 62, 66, 70, 74, 78], "train": [1, 3, 4, 5, 7, 8, 9, 11, 12, 13, 15, 16, 17, 19, 20, 21, 23, 24, 25, 27, 28, 29, 31, 32, 33, 35, 36, 37, 39, 40, 41, 43, 44, 45, 47, 48, 49, 51, 52, 53, 55, 56, 57, 59, 60, 61, 63, 64, 65, 67, 68, 69, 71, 72, 73, 75, 76, 77, 79, 80]}

coco_splits/split_2.json

@@ -0,0 +1 @@
+{"val": [3, 7, 11, 15, 19, 23, 27, 31, 35, 39, 43, 47, 51, 55, 59, 63, 67, 71, 75, 79], "train": [1, 2, 4, 5, 6, 8, 9, 10, 12, 13, 14, 16, 17, 18, 20, 21, 22, 24, 25, 26, 28, 29, 30, 32, 33, 34, 36, 37, 38, 40, 41, 42, 44, 45, 46, 48, 49, 50, 52, 53, 54, 56, 57, 58, 60, 61, 62, 64, 65, 66, 68, 69, 70, 72, 73, 74, 76, 77, 78, 80]}

coco_splits/split_3.json

@@ -0,0 +1 @@
+{"val": [4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, 48, 52, 56, 60, 64, 68, 72, 76, 80], "train": [1, 2, 3, 5, 6, 7, 9, 10, 11, 13, 14, 15, 17, 18, 19, 21, 22, 23, 25, 26, 27, 29, 30, 31, 33, 34, 35, 37, 38, 39, 41, 42, 43, 45, 46, 47, 49, 50, 51, 53, 54, 55, 57, 58, 59, 61, 62, 63, 65, 66, 67, 69, 70, 71, 73, 74, 75, 77, 78, 79]}

config/H_48_D_4_proto.json

@@ -0,0 +1,147 @@
+{
+    "dataset": "coco",
+    "method": "fcn_segmentor",
+    "data": {
+      "image_tool": "cv2",
+      "input_mode": "BGR",
+      "num_classes": 19,
+      "label_list": [7, 8, 11, 12, 13, 17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 31, 32, 33],
+      "data_dir": "/teamspace/studios/this_studio/lessdata",
+      "workers": 1
+    },
+   "train": {
+      "batch_size": 1,
+      "data_transformer": {
+        "size_mode": "fix_size",
+        "input_size": [1024, 512],
+        "align_method": "only_pad",
+        "pad_mode": "random"
+      }
+    },
+    "val": {
+      "batch_size": 1,
+      "mode": "ss_test",
+      "data_transformer": {
+        "size_mode": "fix_size",
+        "input_size": [2048, 1024],
+        "align_method": "only_pad"
+      }
+    },
+    "test": {
+      "batch_size": 1,
+      "mode": "ss_test",
+      "out_dir": "/msravcshare/dataset/seg_result/cityscapes",
+      "data_transformer": {
+        "size_mode": "fix_size",
+        "input_size": [2048, 1024],
+        "align_method": "only_pad"
+      }
+    },
+    "train_trans": {
+      "trans_seq": ["random_resize", "random_crop", "random_hflip", "random_brightness"],
+      "random_brightness": {
+        "ratio": 1.0,
+        "shift_value": 10
+      },
+      "random_hflip": {
+        "ratio": 0.5,
+        "swap_pair": []
+      },
+      "random_resize": {
+        "ratio": 1.0,
+        "method": "random",
+        "scale_range": [0.5, 2.0],
+        "aspect_range": [0.9, 1.1]
+      },
+      "random_crop":{
+        "ratio": 1.0,
+        "crop_size": [1024, 512],
+        "method": "random",
+        "allow_outside_center": false
+      }
+    },
+    "val_trans": {
+      "trans_seq": []
+    },
+    "normalize": {
+      "div_value": 255.0,
+      "mean_value": [0.485, 0.456, 0.406],
+      "mean": [0.485, 0.456, 0.406],
+      "std": [0.229, 0.224, 0.225]
+    },
+    "checkpoints": {
+      "checkpoints_name": "fs_baseocnet_cityscapes_seg",
+      "checkpoints_dir": "./checkpoints/cityscapes",
+      "save_iters": 1000
+    },
+    "network":{
+      "backbone": "deepbase_resnet101_dilated8",
+      "multi_grid": [1, 1, 1],
+      "model_name": "base_ocnet",
+      "bn_type": "torchsyncbn",
+      "stride": 8,
+      "factors": [[8, 8]],
+      "loss_weights": {
+        "corr_loss": 0.01,
+        "aux_loss": 0.4,
+        "seg_loss": 1.0
+      }
+    },
+    "logging": {
+      "logfile_level": "info",
+      "stdout_level": "info",
+      "log_file": "./log/cityscapes/fs_baseocnet_cityscapes_seg.log",
+      "log_format": "%(asctime)s %(levelname)-7s %(message)s",
+      "rewrite": true
+    },
+    "lr": {
+      "base_lr": 0.01,
+      "metric": "iters",
+      "lr_policy": "lambda_poly",
+      "step": {
+        "gamma": 0.5,
+        "step_size": 100
+      }
+    },
+    "solver": {
+      "display_iter": 10,
+      "test_interval": 2000,
+      "max_iters": 40000
+    },
+    "optim": {
+      "optim_method": "sgd",
+      "adam": {
+        "betas": [0.9, 0.999],
+        "eps": 1e-08,
+        "weight_decay": 0.0001
+      },
+      "sgd": {
+        "weight_decay": 0.0005,
+        "momentum": 0.9,
+        "nesterov": false
+      }
+    },
+    "loss": {
+      "loss_type": "pixel_prototype_ce_loss",
+      "params": {
+        "ce_weight": [0.8373, 0.9180, 0.8660, 1.0345, 1.0166, 0.9969, 0.9754,
+                      1.0489, 0.8786, 1.0023, 0.9539, 0.9843, 1.1116, 0.9037,
+                      1.0865, 1.0955, 1.0865, 1.1529, 1.0507],
+        "ce_reduction": "mean",
+        "ce_ignore_index": -1,
+        "ohem_minkeep": 100000,
+        "ohem_thresh": 0.9
+      }
+    },
+    "protoseg": {
+      "gamma": 0.999,
+      "loss_ppc_weight": 0.01,
+      "loss_ppd_weight": 0.001,
+      "num_prototype": 10,
+      "pretrain_prototype": false,
+      "use_rmi": false,
+      "use_prototype": true,
+      "update_prototype": true,
+      "warmup_iters": 0
+    }
+}

config/coco.yaml

@@ -5,7 +5,7 @@ DATA:
   val_list: lists/coco/val.txt
   split: 0
   use_split_coco: True
-  workers: 3
+  workers: 1
   image_size: 417
   mean: [0.485, 0.456, 0.406]
   std: [0.229, 0.224, 0.225]
@@ -24,10 +24,10 @@ EVALUATION:
   ckpt_path: model_ckpt/
   load_model_id: 1
   ckpt_used: model
-  test_num: 10000
-  shot: 5
-  batch_size_val: 50
-  n_runs: 5
+  test_num: 5
+  shot: 2
+  batch_size_val: 1
+  n_runs: 2
   support_only_one_novel: True
   use_training_images_for_supports: False
   generate_new_support_set_for_each_task: False
@@ -41,4 +41,4 @@ CLASSIFIER:
   cls_lr: 0.00125
   pi_estimation_strategy: self
   pi_update_at: [10]
-  fine_tune_base_classifier: True
+  fine_tune_base_classifier: True

config/coco_resnet_base.yaml

@@ -0,0 +1,66 @@
+Data:
+  data_root:  ../data/coco
+  train_list: ./lists/coco/train.txt
+  val_list:  ./lists/coco/val.txt
+  classes: 61
+
+
+Train:
+  # Aug
+  train_h: 417
+  train_w: 417
+  val_size: 417
+  scale_min: 0.5  # minimum random scale
+  scale_max: 2.0 # maximum random scale
+  rotate_min: -10  # minimum random rotate
+  rotate_max: 10  # maximum random rotate
+  ignore_label: 255
+  padding_label: 255
+  # Dataset & Mode
+  split: 0
+  data_set: 'coco'
+  use_split_coco: True # True means FWB setting
+  # Optimizer
+  batch_size: 3 # batch size for training (bs12 for 1GPU)
+  base_lr: 2.5e-4
+  epochs: 20
+  start_epoch: 0
+  stop_interval: 75 # stop when the best result is not updated for "stop_interval" epochs
+  index_split: -1 # index for determining the params group with 10x learning rate
+  power: 0.9 # 0 means no decay
+  momentum: 0.9
+  weight_decay: 0.0001
+  warmup: False
+  # Viz & Save & Resume
+  print_freq: 10
+  save_freq: 5
+  resume: # path to latest checkpoint (default: none, such as epoch_10.pth)  
+  # Validate
+  evaluate: True
+  fix_random_seed_val: True
+  batch_size_val: 4
+  resized_val: True
+  ori_resize: False  # use original label for evaluation
+  # Else
+  workers: 8  # 8 data loader workers
+  manual_seed: 321
+  seed_deterministic: False
+  zoom_factor: 8  # zoom factor for final prediction during training, be in [1, 2, 4, 8]
+
+Method:
+  layers: 50
+  vgg: False
+
+
+
+## deprecated multi-processing training
+# Distributed:
+#   dist_url: tcp://127.0.0.1:6789
+#   dist_backend: 'nccl'
+#   multiprocessing_distributed: False
+#   world_size: 1
+#   rank: 0
+#   use_apex: False
+#   opt_level: 'O0'
+#   keep_batchnorm_fp32:
+#   loss_scale:ls

data/coco/create_masks.py

@@ -5,7 +5,8 @@
 from pycocotools.coco import COCO
 
 for dataset in ['train2014', 'val2014']:
-    annFile = os.path.join('annotations', f'instances_{dataset}.json')
+    annFile = os.path.join('coco-2014/raw', f'instances_{dataset}.json')
+    print(f'PATH...... {annFile}')
     img_dir = dataset
     save_dir = 'train' if 'train' in dataset else 'val'
 

downloadCoco.py

@@ -0,0 +1,9 @@
+import fiftyone as fo
+import fiftyone.zoo as foz
+
+# To download the COCO dataset for only the "person" and "car" classes
+dataset = foz.load_zoo_dataset(
+    "coco-2014",
+    splits=["train", "validation", "test"],
+    label_types=["detections", "segmentations"]
+)

model/ASPP.py

@@ -0,0 +1,49 @@
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.utils.data
+
+class ASPP(nn.Module):
+    def __init__(self, out_channels=256):
+        super(ASPP, self).__init__()
+        self.layer6_0 = nn.Sequential(
+            nn.Conv2d(out_channels , out_channels, kernel_size=1, stride=1, padding=0, bias=True),
+            nn.ReLU(),
+        )
+        self.layer6_1 = nn.Sequential(
+            nn.Conv2d(out_channels , out_channels, kernel_size=1, stride=1, padding=0, bias=True),
+            nn.ReLU(),
+            )
+        self.layer6_2 = nn.Sequential(
+            nn.Conv2d(out_channels , out_channels , kernel_size=3, stride=1, padding=6,dilation=6, bias=True),
+            nn.ReLU(),
+            )
+        self.layer6_3 = nn.Sequential(
+            nn.Conv2d(out_channels , out_channels, kernel_size=3, stride=1, padding=12, dilation=12, bias=True),
+            nn.ReLU(),
+            )
+        self.layer6_4 = nn.Sequential(
+            nn.Conv2d(out_channels , out_channels , kernel_size=3, stride=1, padding=18, dilation=18, bias=True),
+            nn.ReLU(),
+            )
+
+        self._init_weight()
+
+    def _init_weight(self):
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                torch.nn.init.kaiming_normal_(m.weight)
+            elif isinstance(m, nn.BatchNorm2d):
+                m.weight.data.fill_(1)
+                m.bias.data.zero_()
+
+    def forward(self, x):
+        feature_size = x.shape[-2:]
+        global_feature = F.avg_pool2d(x, kernel_size=feature_size)
+
+        global_feature = self.layer6_0(global_feature)
+
+        global_feature = global_feature.expand(-1, -1, feature_size[0], feature_size[1])
+        out = torch.cat(
+            [global_feature, self.layer6_1(x), self.layer6_2(x), self.layer6_3(x), self.layer6_4(x)], dim=1)
+        return out