Merge pull request #786 from RainFrost1/develop_reg

修改Neck模块，添加车辆任务的完整训练代码

ppcls/arch/__init__.py

-#copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+#copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
 #
 #Licensed under the Apache License, Version 2.0 (the "License");
 #you may not use this file except in compliance with the License.
@@ -17,11 +17,9 @@ import importlib
 
 import paddle.nn as nn
 
-from . import backbone
-from . import gears
-
+from . import backbone, gears
 from .backbone import *
-from .gears import *
+from .gears import build_gear
 from .utils import *
 
 __all__ = ["build_model", "RecModel"]
@@ -38,34 +36,28 @@ def build_model(config):
 class RecModel(nn.Layer):
     def __init__(self, **config):
         super().__init__()
-
         backbone_config = config["Backbone"]
         backbone_name = backbone_config.pop("name")
         self.backbone = eval(backbone_name)(**backbone_config)
+        if "BackboneStopLayer" in config:
+            backbone_stop_layer = config["BackboneStopLayer"]["name"]
+            self.backbone.stop_after(backbone_stop_layer)
 
-        assert "Stoplayer" in config, "Stoplayer should be specified in retrieval task \
-                please specified a Stoplayer config"
-
-        stop_layer_config = config["Stoplayer"]
-        self.backbone.stop_after(stop_layer_config["name"])
-
-        if stop_layer_config.get("embedding_size", 0) > 0:
-            self.neck = nn.Linear(stop_layer_config["output_dim"],
-                                  stop_layer_config["embedding_size"])
-            embedding_size = stop_layer_config["embedding_size"]
+        if "Neck" in config:
+            self.neck = build_gear(config["Neck"])
         else:
             self.neck = None
-            embedding_size = stop_layer_config["output_dim"]
-
-        assert "Head" in config, "Head should be specified in retrieval task \
-                please specify a Head config"
 
-        config["Head"]["embedding_size"] = embedding_size
-        self.head = build_head(config["Head"])
+        if "Head" in config:
+            self.head = build_gear(config["Head"])
+        else:
+            self.head = None
 
     def forward(self, x, label):
         x = self.backbone(x)
         if self.neck is not None:
             x = self.neck(x)
-        y = self.head(x, label)
+        y = x
+        if self.head is not None:
+            y = self.head(x, label)
         return {"features": x, "logits": y}

ppcls/arch/gears/__init__.py

@@ -16,12 +16,15 @@ from .arcmargin import ArcMargin
 from .cosmargin import CosMargin
 from .circlemargin import CircleMargin
 from .fc import FC
+from .vehicle_neck import VehicleNeck
 
-__all__ = ['build_head']
+__all__ = ['build_gear']
 
 
-def build_head(config):
-    support_dict = ['ArcMargin', 'CosMargin', 'CircleMargin', 'FC']
+def build_gear(config):
+    support_dict = [
+        'ArcMargin', 'CosMargin', 'CircleMargin', 'FC', 'VehicleNeck'
+    ]
     module_name = config.pop('name')
     assert module_name in support_dict, Exception(
         'head only support {}'.format(support_dict))

ppcls/arch/gears/vehicle_neck.py

+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import, division, print_function
+
+import paddle
+import paddle.nn as nn
+
+
+class VehicleNeck(nn.Layer):
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 kernel_size=1,
+                 stride=1,
+                 padding=0,
+                 dilation=1,
+                 groups=1,
+                 padding_mode='zeros',
+                 weight_attr=None,
+                 bias_attr=None,
+                 data_format='NCHW'):
+        super().__init__()
+        self.conv = nn.Conv2D(
+            in_channels=in_channels,
+            out_channels=out_channels,
+            kernel_size=kernel_size,
+            stride=stride,
+            padding=padding,
+            dilation=dilation,
+            groups=groups,
+            padding_mode=padding_mode,
+            weight_attr=weight_attr,
+            bias_attr=weight_attr,
+            data_format=data_format)
+        self.flatten = nn.Flatten()
+
+    def forward(self, x):
+        x = self.conv(x)
+        x = self.flatten(x)
+        return x

ppcls/configs/Vehicle/ResNet50.yaml

@@ -19,11 +19,14 @@ Global:
 Arch:
   name: "RecModel"
   Backbone: 
-    name: "ResNet50"
-  Stoplayer: 
-    name: "flatten_0" 
-    output_dim: 2048
-    embedding_size: 512
+    name: "ResNet50_last_stage_stride1"
+    pretrained: True
+  BackboneStopLayer:
+    name: "adaptive_avg_pool2d_0"
+  Neck:
+    name: "VehicleNeck"
+    in_channels: 2048
+    out_channels: 512
   Head:
     name: "ArcMargin"  
     embedding_size: 512
@@ -88,7 +91,7 @@ DataLoader:
 
     sampler:
         name: DistributedRandomIdentitySampler
-        batch_size: 64
+        batch_size: 128
         num_instances: 2
         drop_last: False
         shuffle: True
@@ -114,7 +117,7 @@ DataLoader:
               order: ''
     sampler:
         name: DistributedBatchSampler
-        batch_size: 64
+        batch_size: 128
         drop_last: False
         shuffle: False
     loader:

ppcls/configs/Vehicle/ResNet50_ReID.yaml

@@ -19,16 +19,18 @@ Global:
   num_split: 1
   feature_normalize: True
 
-
 # model architecture
 Arch:
   name: "RecModel"
   Backbone: 
-    name: "ResNet50"
-  Stoplayer: 
-    name: "flatten_0" 
-    output_dim: 2048
-    embedding_size: 512
+    name: "ResNet50_last_stage_stride1"
+    pretrained: True
+  BackboneStopLayer:
+    name: "adaptive_avg_pool2d_0"
+  Neck:
+    name: "VehicleNeck"
+    in_channels: 2048
+    out_channels: 512
   Head:
     name: "ArcMargin"  
     embedding_size: 512
@@ -41,9 +43,9 @@ Loss:
   Train:
     - CELoss:
         weight: 1.0
-    - TripletLossV2:
+    - SupConLoss:
         weight: 1.0
-        margin: 0.5
+        views: 2
   Eval:
     - CELoss:
         weight: 1.0
@@ -68,7 +70,7 @@ DataLoader:
     dataset:
         name: "VeriWild"
         image_root: "/work/dataset/VeRI-Wild/images/"
-        cls_label_path: "/work/dataset/VeRI-Wild/train_test_split/debug_train.txt"
+        cls_label_path: "/work/dataset/VeRI-Wild/train_test_split/train_list_start0.txt"
         transform_ops:
           - ResizeImage:
               size: 224
@@ -103,7 +105,7 @@ DataLoader:
     dataset: 
         name: "VeriWild"
         image_root: "/work/dataset/VeRI-Wild/images"
-        cls_label_path: "/work/dataset/VeRI-Wild/train_test_split/debug_test_query.txt"
+        cls_label_path: "/work/dataset/VeRI-Wild/train_test_split/test_3000_id_query.txt"
         transform_ops:
           - ResizeImage:
               size: 224
@@ -126,7 +128,7 @@ DataLoader:
     dataset: 
         name: "VeriWild"
         image_root: "/work/dataset/VeRI-Wild/images"
-        cls_label_path: "/work/dataset/VeRI-Wild/train_test_split/debug_test.txt"
+        cls_label_path: "/work/dataset/VeRI-Wild/train_test_split/test_3000_id.txt"
         transform_ops:
           - ResizeImage:
               size: 224

ppcls/loss/__init__.py

@@ -11,9 +11,11 @@ from .msmloss import MSMLoss
 from .npairsloss import NpairsLoss
 from .trihardloss import TriHardLoss
 from .triplet import TripletLoss, TripletLossV2
+from .supconloss import SupConLoss
 from .pairwisecosface import PairwiseCosface
 
 
+
 class CombinedLoss(nn.Layer):
     def __init__(self, config_list):
         super().__init__()

ppcls/loss/supconloss.py

+import paddle
+from paddle import nn
+
+
+class SupConLoss(nn.Layer):
+    """Supervised Contrastive Learning: https://arxiv.org/pdf/2004.11362.pdf.
+    It also supports the unsupervised contrastive loss in SimCLR"""
+
+    def __init__(self,
+                 views=16,
+                 temperature=0.07,
+                 contrast_mode='all',
+                 base_temperature=0.07,
+                 normalize_feature=True):
+        super(SupConLoss, self).__init__()
+        self.temperature = paddle.to_tensor(temperature)
+        self.contrast_mode = contrast_mode
+        self.base_temperature = paddle.to_tensor(base_temperature)
+        self.num_ids = None
+        self.views = views
+        self.normalize_feature = normalize_feature
+
+    def forward(self, features, labels, mask=None):
+        """Compute loss for model. If both `labels` and `mask` are None,
+        it degenerates to SimCLR unsupervised loss:
+        https://arxiv.org/pdf/2002.05709.pdf
+        Args:
+            features: hidden vector of shape [bsz, n_views, ...].
+            labels: ground truth of shape [bsz].
+            mask: contrastive mask of shape [bsz, bsz], mask_{i,j}=1 if sample j
+                has the same class as sample i. Can be asymmetric.
+        Returns:
+            A loss scalar.
+        """
+        features = features["features"]
+        if self.num_ids is None:
+            self.num_ids = int(features.shape[0] / self.views)
+
+        if self.normalize_feature:
+            features = 1. * features / (paddle.expand_as(
+                paddle.norm(
+                    features, p=2, axis=-1, keepdim=True), features) + 1e-12)
+        features = features.reshape([self.num_ids, self.views, -1])
+        labels = labels.reshape([self.num_ids, self.views])[:, 0]
+
+        if len(features.shape) < 3:
+            raise ValueError('`features` needs to be [bsz, n_views, ...],'
+                             'at least 3 dimensions are required')
+        if len(features.shape) > 3:
+            features = features.reshape(
+                [features.shape[0], features.shape[1], -1])
+
+        batch_size = features.shape[0]
+        if labels is not None and mask is not None:
+            raise ValueError('Cannot define both `labels` and `mask`')
+        elif labels is None and mask is None:
+            mask = paddle.eye(batch_size, dtype='float32')
+        elif labels is not None:
+            labels = labels.reshape([-1, 1])
+            if labels.shape[0] != batch_size:
+                raise ValueError(
+                    'Num of labels does not match num of features')
+            mask = paddle.cast(
+                paddle.equal(labels, paddle.t(labels)), 'float32')
+        else:
+            mask = paddle.cast(mask, 'float32')
+
+        contrast_count = features.shape[1]
+        contrast_feature = paddle.concat(
+            paddle.unbind(
+                features, axis=1), axis=0)
+        if self.contrast_mode == 'one':
+            anchor_feature = features[:, 0]
+            anchor_count = 1
+        elif self.contrast_mode == 'all':
+            anchor_feature = contrast_feature
+            anchor_count = contrast_count
+        else:
+            raise ValueError('Unknown mode: {}'.format(self.contrast_mode))
+
+        # compute logits
+        anchor_dot_contrast = paddle.divide(
+            paddle.matmul(anchor_feature, paddle.t(contrast_feature)),
+            self.temperature)
+        # for numerical stability
+        logits_max = paddle.max(anchor_dot_contrast, axis=1, keepdim=True)
+        logits = anchor_dot_contrast - logits_max.detach()
+
+        # tile mask
+        mask = paddle.tile(mask, [anchor_count, contrast_count])
+
+        logits_mask = 1 - paddle.eye(batch_size * anchor_count)
+        mask = mask * logits_mask
+
+        # compute log_prob
+        exp_logits = paddle.exp(logits) * logits_mask
+        log_prob = logits - paddle.log(
+            paddle.sum(exp_logits, axis=1, keepdim=True))
+
+        # compute mean of log-likelihood over positive
+        mean_log_prob_pos = paddle.sum((mask * log_prob),
+                                       axis=1) / paddle.sum(mask, axis=1)
+
+        # loss
+        loss = -(self.temperature / self.base_temperature) * mean_log_prob_pos
+        loss = paddle.mean(loss.reshape([anchor_count, batch_size]))
+
+        return {"SupConLoss": loss}