Minor changes

Some minor changes, such as class name changing, remove extra blank line, etc.

demo/README.md

@@ -2,18 +2,9 @@
 
 We provide a command line tool to run a simple demo of builtin models.
 
-You can run this command to get rank visualization results by cosine similarites between different images.
+You can run this command to get cosine similarites between different images 
 
-```shell script
-python3 demo/visualize_result.py --config-file logs/dukemtmc/mgn_R50-ibn/config.yaml \
---parallel --vis-label --dataset-name 'DukeMTMC' --output logs/mgn_duke_vis \
---opts MODEL.WEIGHTS logs/dukemtmc/mgn_R50-ibn/model_final.pth
-```
-
-You can also run this command to extract image features.
-
-```shell script
-python3 demo/demo.py --config-file logs/dukemtmc/sbs_R50/config.yaml \
---parallel --input tools/deploy/test_data/*.jpg --output sbs_R50_feat \
---opts MODEL.WEIGHTS logs/dukemtmc/sbs_R50/model_final.pth
+```bash
+cd demo/
+sh run_demo.sh
 ```

demo/demo.py

@@ -9,6 +9,7 @@ import glob
 import os
 import sys
 
+import torch.nn.functional as F
 import cv2
 import numpy as np
 import tqdm
@@ -23,7 +24,7 @@ from fastreid.utils.file_io import PathManager
 from predictor import FeatureExtractionDemo
 
 # import some modules added in project like this below
-# sys.path.append('../projects/PartialReID')
+# sys.path.append("projects/PartialReID")
 # from partialreid import *
 
 cudnn.benchmark = True
@@ -72,6 +73,13 @@ def get_parser():
     return parser
 
 
+def postprocess(features):
+    # Normalize feature to compute cosine distance
+    features = F.normalize(features)
+    features = features.cpu().data.numpy()
+    return features
+
+
 if __name__ == '__main__':
     args = get_parser().parse_args()
     cfg = setup_cfg(args)
@@ -85,5 +93,5 @@ if __name__ == '__main__':
         for path in tqdm.tqdm(args.input):
             img = cv2.imread(path)
             feat = demo.run_on_image(img)
-            feat = feat.numpy()
+            feat = postprocess(feat)
             np.save(os.path.join(args.output, os.path.basename(path).split('.')[0] + '.npy'), feat)

fastreid/data/transforms/build.py

@@ -78,8 +78,8 @@ def build_transforms(cfg, is_train=True):
         if do_cj:
             res.append(T.RandomApply([T.ColorJitter(cj_brightness, cj_contrast, cj_saturation, cj_hue)], p=cj_prob))
         if do_affine:
-            res.append(T.RandomAffine(degrees=0, translate=None, scale=[0.9, 1.1], shear=None, resample=False,
-                                      fillcolor=128))
+            res.append(T.RandomAffine(degrees=10, translate=None, scale=[0.9, 1.1], shear=0.1, resample=False,
+                                      fillcolor=0))
         if do_augmix:
             res.append(AugMix(prob=augmix_prob))
         res.append(ToTensor())

fastreid/layers/__init__.py

@@ -5,11 +5,15 @@
 """
 
 from .activation import *
-from .batch_drop import BatchDrop
 from .batch_norm import *
 from .context_block import ContextBlock
+from .drop import DropPath, DropBlock2d, drop_block_2d, drop_path
 from .frn import FRN, TLU
+from .gather_layer import GatherLayer
+from .helpers import to_ntuple, to_2tuple, to_3tuple, to_4tuple, make_divisible
 from .non_local import Non_local
 from .se_layer import SELayer
 from .splat import SplAtConv2d, DropBlock2D
-from .gather_layer import GatherLayer
+from .weight_init import (
+    trunc_normal_, variance_scaling_, lecun_normal_, weights_init_kaiming, weights_init_classifier
+)

fastreid/layers/any_softmax.py

@@ -23,7 +23,7 @@ class Linear(nn.Module):
         self.m = margin
 
     def forward(self, logits, targets):
-        return logits
+        return logits.mul_(self.s)
 
     def extra_repr(self):
         return f"num_classes={self.num_classes}, scale={self.s}, margin={self.m}"

fastreid/layers/batch_drop.py

@@ -1,32 +0,0 @@
-# encoding: utf-8
-"""
-@author:  liaoxingyu
-@contact: sherlockliao01@gmail.com
-"""
-
-import random
-
-from torch import nn
-
-
-class BatchDrop(nn.Module):
-    """ref: https://github.com/daizuozhuo/batch-dropblock-network/blob/master/models/networks.py
-    batch drop mask
-    """
-
-    def __init__(self, h_ratio, w_ratio):
-        super(BatchDrop, self).__init__()
-        self.h_ratio = h_ratio
-        self.w_ratio = w_ratio
-
-    def forward(self, x):
-        if self.training:
-            h, w = x.size()[-2:]
-            rh = round(self.h_ratio * h)
-            rw = round(self.w_ratio * w)
-            sx = random.randint(0, h - rh)
-            sy = random.randint(0, w - rw)
-            mask = x.new_ones(x.size())
-            mask[:, :, sx:sx + rh, sy:sy + rw] = 0
-            x = x * mask
-        return x

fastreid/layers/pooling.py

@@ -61,7 +61,7 @@ class GeneralizedMeanPooling(nn.Module):
                      be the same as that of the input.
     """
 
-    def __init__(self, norm=3, output_size=1, eps=1e-6, *args, **kwargs):
+    def __init__(self, norm=3, output_size=(1, 1), eps=1e-6, *args, **kwargs):
         super(GeneralizedMeanPooling, self).__init__()
         assert norm > 0
         self.p = float(norm)
@@ -82,7 +82,7 @@ class GeneralizedMeanPoolingP(GeneralizedMeanPooling):
     """ Same, but norm is trainable
     """
 
-    def __init__(self, norm=3, output_size=1, eps=1e-6, *args, **kwargs):
+    def __init__(self, norm=3, output_size=(1, 1), eps=1e-6, *args, **kwargs):
         super(GeneralizedMeanPoolingP, self).__init__(norm, output_size, eps)
         self.p = nn.Parameter(torch.ones(1) * norm)
 

fastreid/modeling/losses/triplet_loss.py

@@ -42,7 +42,7 @@ def hard_example_mining(dist_mat, is_pos, is_neg):
     dist_ap, _ = torch.max(dist_mat * is_pos, dim=1)
     # `dist_an` means distance(anchor, negative)
     # both `dist_an` and `relative_n_inds` with shape [N]
-    dist_an, _ = torch.min(dist_mat * is_neg + is_pos * 99999999., dim=1)
+    dist_an, _ = torch.min(dist_mat * is_neg + is_pos * 1e9, dim=1)
 
     return dist_ap, dist_an
 

projects/FastAttr/fastattr/modeling/attr_head.py

@@ -10,7 +10,7 @@ from torch import nn
 
 from fastreid.modeling.heads import EmbeddingHead
 from fastreid.modeling.heads.build import REID_HEADS_REGISTRY
-from fastreid.utils.weight_init import weights_init_kaiming
+from fastreid.layers.weight_init import weights_init_kaiming
 
 
 @REID_HEADS_REGISTRY.register()

projects/FastClas/fastclas/__init__.py

@@ -5,4 +5,6 @@
 """
 
 from .bee_ant import *
+from .distracted_driver import *
 from .dataset import ClasDataset
+from .trainer import ClasTrainer

projects/FastClas/fastclas/bee_ant.py

@@ -10,6 +10,7 @@ import os
 from fastreid.data.datasets import DATASET_REGISTRY
 from fastreid.data.datasets.bases import ImageDataset
 
+
 __all__ = ["Hymenoptera"]
 
 

projects/FastClas/fastclas/dataset.py

@@ -12,18 +12,22 @@ from fastreid.data.data_utils import read_image
 class ClasDataset(Dataset):
     """Image Person ReID Dataset"""
 
-    def __init__(self, img_items, transform=None):
+    def __init__(self, img_items, transform=None, idx_to_class=None):
         self.img_items = img_items
         self.transform = transform
 
-        classes = set()
-        for i in img_items:
-            classes.add(i[1])
+        if idx_to_class is not None:
+            self.idx_to_class = idx_to_class
+            self.class_to_idx = {clas_name: int(i) for i, clas_name in self.idx_to_class.items()}
+            self.classes = sorted(list(self.idx_to_class.values()))
+        else:
+            classes = set()
+            for i in img_items:
+                classes.add(i[1])
 
-        self.classes = list(classes)
-        self.classes.sort()
-        self.class_to_idx = {cls_name: i for i, cls_name in enumerate(self.classes)}
-        self.idx_to_class = {idx: clas for clas, idx in self.class_to_idx.items()}
+            self.classes = sorted(list(classes))
+            self.class_to_idx = {cls_name: i for i, cls_name in enumerate(self.classes)}
+            self.idx_to_class = {idx: clas for clas, idx in self.class_to_idx.items()}
 
     def __len__(self):
         return len(self.img_items)

projects/FastClas/fastclas/trainer.py

@@ -0,0 +1,82 @@
+# encoding: utf-8
+"""
+@author:  xingyu liao
+@contact: sherlockliao01@gmail.com
+"""
+
+import json
+import logging
+import os
+
+from fastreid.data.build import _root
+from fastreid.data.build import build_reid_train_loader, build_reid_test_loader
+from fastreid.data.datasets import DATASET_REGISTRY
+from fastreid.data.transforms import build_transforms
+from fastreid.engine import DefaultTrainer
+from fastreid.evaluation.clas_evaluator import ClasEvaluator
+from fastreid.utils import comm
+from fastreid.utils.checkpoint import PathManager
+from .dataset import ClasDataset
+
+
+class ClasTrainer(DefaultTrainer):
+    idx2class = None
+
+    @classmethod
+    def build_train_loader(cls, cfg):
+        """
+        Returns:
+            iterable
+        It now calls :func:`fastreid.data.build_reid_train_loader`.
+        Overwrite it if you'd like a different data loader.
+        """
+        logger = logging.getLogger("fastreid.clas_dataset")
+        logger.info("Prepare training set")
+
+        train_items = list()
+        for d in cfg.DATASETS.NAMES:
+            data = DATASET_REGISTRY.get(d)(root=_root)
+            if comm.is_main_process():
+                data.show_train()
+            train_items.extend(data.train)
+        transforms = build_transforms(cfg, is_train=True)
+        train_set = ClasDataset(train_items, transforms)
+        cls.idx2class = train_set.idx_to_class
+
+        data_loader = build_reid_train_loader(cfg, train_set=train_set)
+        return data_loader
+
+    @classmethod
+    def build_test_loader(cls, cfg, dataset_name):
+        """
+        Returns:
+            iterable
+        It now calls :func:`fastreid.data.build_reid_test_loader`.
+        Overwrite it if you'd like a different data loader.
+        """
+        data = DATASET_REGISTRY.get(dataset_name)(root=_root)
+        if comm.is_main_process():
+            data.show_test()
+        transforms = build_transforms(cfg, is_train=False)
+
+        test_set = ClasDataset(data.query, transforms, cls.idx2class)
+        data_loader, _ = build_reid_test_loader(cfg, test_set=test_set)
+        return data_loader
+
+    @classmethod
+    def build_evaluator(cls, cfg, dataset_name, output_dir=None):
+        data_loader = cls.build_test_loader(cfg, dataset_name)
+        return data_loader, ClasEvaluator(cfg, output_dir)
+
+    @staticmethod
+    def auto_scale_hyperparams(cfg, num_classes):
+        cfg = DefaultTrainer.auto_scale_hyperparams(cfg, num_classes)
+
+        # Save index to class dictionary
+        output_dir = cfg.OUTPUT_DIR
+        if comm.is_main_process() and output_dir:
+            path = os.path.join(output_dir, "idx2class.json")
+            with PathManager.open(path, "w") as f:
+                json.dump(ClasTrainer.idx2class, f)
+
+        return cfg

projects/FastClas/train_net.py

@@ -14,75 +14,11 @@ sys.path.append('.')
 
 from fastreid.config import get_cfg
 from fastreid.engine import default_argument_parser, default_setup, launch
-from fastreid.data.build import build_reid_train_loader, build_reid_test_loader
-from fastreid.evaluation.clas_evaluator import ClasEvaluator
 from fastreid.utils.checkpoint import Checkpointer, PathManager
-from fastreid.utils import comm
-from fastreid.engine import DefaultTrainer
-from fastreid.data.datasets import DATASET_REGISTRY
-from fastreid.data.transforms import build_transforms
-from fastreid.data.build import _root
 
 from fastclas import *
 
 
-class ClasTrainer(DefaultTrainer):
-
-    @classmethod
-    def build_train_loader(cls, cfg):
-        """
-        Returns:
-            iterable
-        It now calls :func:`fastreid.data.build_reid_train_loader`.
-        Overwrite it if you'd like a different data loader.
-        """
-        logger = logging.getLogger("fastreid.clas_dataset")
-        logger.info("Prepare training set")
-
-        train_items = list()
-        for d in cfg.DATASETS.NAMES:
-            data = DATASET_REGISTRY.get(d)(root=_root)
-            if comm.is_main_process():
-                data.show_train()
-            train_items.extend(data.train)
-
-        transforms = build_transforms(cfg, is_train=True)
-        train_set = ClasDataset(train_items, transforms)
-
-        data_loader = build_reid_train_loader(cfg, train_set=train_set)
-
-        # Save index to class dictionary
-        output_dir = cfg.OUTPUT_DIR
-        if comm.is_main_process() and output_dir:
-            path = os.path.join(output_dir, "idx2class.json")
-            with PathManager.open(path, "w") as f:
-                json.dump(train_set.idx_to_class, f)
-
-        return data_loader
-
-    @classmethod
-    def build_test_loader(cls, cfg, dataset_name):
-        """
-        Returns:
-            iterable
-        It now calls :func:`fastreid.data.build_reid_test_loader`.
-        Overwrite it if you'd like a different data loader.
-        """
-
-        data = DATASET_REGISTRY.get(dataset_name)(root=_root)
-        if comm.is_main_process():
-            data.show_test()
-        transforms = build_transforms(cfg, is_train=False)
-        test_set = ClasDataset(data.query, transforms)
-        data_loader, _ = build_reid_test_loader(cfg, test_set=test_set)
-        return data_loader
-
-    @classmethod
-    def build_evaluator(cls, cfg, dataset_name, output_dir=None):
-        data_loader = cls.build_test_loader(cfg, dataset_name)
-        return data_loader, ClasEvaluator(cfg, output_dir)
-
-
 def setup(args):
     """
     Create configs and perform basic setups.
@@ -105,6 +41,16 @@ def main(args):
 
         Checkpointer(model).load(cfg.MODEL.WEIGHTS)  # load trained model
 
+        try:
+            output_dir = os.path.dirname(cfg.MODEL.WEIGHTS)
+            path = os.path.join(output_dir, "idx2class.json")
+            with PathManager.open(path, 'r') as f:
+                idx2class = json.load(f)
+            ClasTrainer.idx2class = idx2class
+        except:
+            logger = logging.getLogger("fastreid.fastclas")
+            logger.info(f"Cannot find idx2class dict in {os.path.dirname(cfg.MODEL.WEIGHTS)}")
+
         res = ClasTrainer.test(cfg, model)
         return res
 

projects/FastTune/configs/search_trial.yml

@@ -55,9 +55,6 @@ INPUT:
   PADDING:
     ENABLED: True
 
-  FLIP:
-    ENABLED: True
-
 DATALOADER:
   SAMPLER_TRAIN: NaiveIdentitySampler
   NUM_INSTANCE: 16

projects/PartialReID/configs/partial_market.yml

@@ -26,7 +26,7 @@ MODEL:
     TRI:
       MARGIN: 0.3
       SCALE: 1.0
-      HARD_MINING: True
+      HARD_MINING: False
 
 DATASETS:
   NAMES: ("Market1501",)
@@ -44,7 +44,6 @@ DATALOADER:
   NUM_INSTANCE: 4
   NUM_WORKERS: 8
 
-  
 SOLVER:
   AMP:
     ENABLED: False
@@ -71,4 +70,4 @@ TEST:
 
 CUDNN_BENCHMARK: True
 
-OUTPUT_DIR: "projects/PartialReID/logs/test_partial"
+OUTPUT_DIR: projects/PartialReID/logs/test_partial

projects/PartialReID/partialreid/dsr_head.py

@@ -11,7 +11,7 @@ from torch import nn
 from fastreid.layers import *
 from fastreid.modeling.heads import EmbeddingHead
 from fastreid.modeling.heads.build import REID_HEADS_REGISTRY
-from fastreid.utils.weight_init import weights_init_kaiming
+from fastreid.layers.weight_init import weights_init_kaiming
 
 
 class OcclusionUnit(nn.Module):

tools/deploy/onnx_export.py

@@ -28,7 +28,6 @@ from fastreid.utils.logger import setup_logger
 # sys.path.append("projects/FastDistill")
 # from fastdistill import *
 
-
 setup_logger(name="fastreid")
 logger = logging.getLogger("fastreid.onnx_export")