update code

2022-04-22 15:52:16 +08:00 · 2022-04-22 15:52:16 +08:00 · 15a0afc67c
parent bc6e7f5cf5
commit 15a0afc67c
6 changed files with 54 additions and 224 deletions
--- a/ppcls/arch/backbone/legendary_models/resnet.py
+++ b/ppcls/arch/backbone/legendary_models/resnet.py
@ -281,7 +281,6 @@ class ResNet(TheseusLayer):
                 lr_mult_list=[1.0, 1.0, 1.0, 1.0, 1.0],
                 data_format="NCHW",
                 input_image_channel=3,
-                 stem_act="relu",
                 return_patterns=None,
                 return_stages=None):
        super().__init__()
--- a/ppcls/configs/Pedestrian/strong_baseline_baseline.yaml
+++ b/ppcls/configs/Pedestrian/strong_baseline_baseline.yaml
@ -6,14 +6,14 @@ Global:
  # pretrained_model: "./pd_model_trace/ISE/ISE_MS_model" # pretrained ISE model for MSMT17
  output_dir: "./output/"
  device: "gpu"
-  save_interval: 10
+  save_interval: 40
  eval_during_train: True
  eval_interval: 10
  epochs: 120
-  print_batch_step: 10
+  print_batch_step: 20
  use_visualdl: False
  # used for static mode and model export
-  image_shape: [3, 128, 256]
+  image_shape: [3, 256, 128]
  save_inference_dir: "./inference"
  eval_mode: "retrieval"

@ -22,14 +22,14 @@ Arch:
  name: "RecModel"
  infer_output_key: "features"
  infer_add_softmax: False
-  Backbone: 
+  Backbone:
    name: "ResNet50_last_stage_stride1"
    pretrained: True
    stem_act: null
  BackboneStopLayer:
    name: "flatten"
  Head:
-    name: "FC"  
+    name: "FC"
    embedding_size: 2048
    class_num: 751
 # loss function config for traing/eval process
@ -138,4 +138,3 @@ Metric:
    - Recallk:
        topk: [1, 5]
    - mAP: {}
-
--- a/ppcls/configs/Pedestrian/strong_baseline_m1.yaml
+++ b/ppcls/configs/Pedestrian/strong_baseline_m1.yaml
@ -4,17 +4,19 @@ Global:
  pretrained_model: null
  output_dir: "./output/"
  device: "gpu"
-  save_interval: 10
+  save_interval: 40
  eval_during_train: True
  eval_interval: 10
  epochs: 120
  print_batch_step: 20
  use_visualdl: False
+  warmup_by_epoch: True
+  eval_mode: "retrieval"
+  re_ranking: False
+  feat_from: "neck" # 'backbone' or 'neck'
  # used for static mode and model export
  image_shape: [3, 256, 128]
  save_inference_dir: "./inference"
-  eval_mode: "retrieval"
-  feat_from: "neck" # 'backbone' or 'neck'

 # model architecture
 Arch:
@ -29,13 +31,17 @@ Arch:
    name: "flatten"
  Neck:
    name: BNNeck
-    num_filters: 2048
-    # trainable: False # TODO: free bn.bias
+    num_features: &feat_dim 2048
+    # trainable: False # TODO: freeze bn.bias
  Head:
    name: "FC"
-    embedding_size: 2048
-    class_num: 751
-    bias_attr: false
+    embedding_size: *feat_dim
+    class_num: &class_num 751
+    weight_attr:
+      initializer:
+        name: Normal
+        std: 0.001
+    bias_attr: False

 # loss function config for traing/eval process
 Loss:
@ -160,4 +166,3 @@ Metric:
    - Recallk:
        topk: [1, 5]
    - mAP: {}
-
--- a/ppcls/configs/Pedestrian/strong_baseline_m1_centerloss.yaml
+++ b/ppcls/configs/Pedestrian/strong_baseline_m1_centerloss.yaml
@ -4,17 +4,19 @@ Global:
  pretrained_model: null
  output_dir: "./output/"
  device: "gpu"
-  save_interval: 10
+  save_interval: 40
  eval_during_train: True
  eval_interval: 10
  epochs: 120
  print_batch_step: 20
  use_visualdl: False
+  warmup_by_epoch: True
+  eval_mode: "retrieval"
+  re_ranking: False
+  feat_from: "neck" # 'backbone' or 'neck'
  # used for static mode and model export
  image_shape: [3, 256, 128]
  save_inference_dir: "./inference"
-  eval_mode: "retrieval"
-  feat_from: "neck" # 'backbone' or 'neck'

 # model architecture
 Arch:
@ -29,13 +31,17 @@ Arch:
    name: "flatten"
  Neck:
    name: BNNeck
-    num_filters: 2048
-    # trainable: False # TODO: free bn.bias
+    num_features: &feat_dim 2048
+    # trainable: False # TODO: freeze bn.bias
  Head:
    name: "FC"
-    embedding_size: 2048
-    class_num: 751
-    bias_attr: false
+    embedding_size: *feat_dim
+    class_num: &class_num 751
+    weight_attr:
+      initializer:
+        name: Normal
+        std: 0.001
+    bias_attr: False

 # loss function config for traing/eval process
 Loss:
@ -43,35 +49,36 @@ Loss:
    - CELoss:
        weight: 1.0
        epsilon: 0.1
-    - TripletLossV2:
+    - TripletLossV3:
        weight: 1.0
        margin: 0.3
        normalize_feature: false
    - CenterLoss:
        weight: 0.0005
-        num_classes: 751
-        feat_dim: 2048
+        num_classes: *class_num
+        feat_dim: *feat_dim
  Eval:
    - CELoss:
        weight: 1.0

 Optimizer:
-  model:
-    name: Adam
-    lr:
-      name: Piecewise
-      decay_epochs: [30, 60]
-      values: [0.00035, 0.000035, 0.0000035]
-      warmup_epoch: 10
-      warmup_start_lr: 0.0000035
-    regularizer:
-      name: 'L2'
-      coeff: 0.0005
-  loss:
-    name: SGD
-    lr:
-      name: Constant
-      learning_rate: 0.5
+  - Adam:
+      scope: model
+      lr:
+        name: Piecewise
+        decay_epochs: [30, 60]
+        values: [0.00035, 0.000035, 0.0000035]
+        warmup_epoch: 10
+        warmup_start_lr: 0.0000035
+        warmup_by_epoch: True
+      regularizer:
+        name: 'L2'
+        coeff: 0.0005
+  - SGD:
+      scope: CenterLoss
+      lr:
+        name: Constant
+        learning_rate: 1000.0 # set to ori_lr*(1/centerloss_weight) to void manually scaling centers' gradidents.

 # data loader for train and eval
 DataLoader:
@ -170,4 +177,3 @@ Metric:
    - Recallk:
        topk: [1, 5]
    - mAP: {}
-
--- a/ppcls/configs/PersonReID/ResNet50_strong_baseline_market1501.yaml
+++ b/ppcls/configs/PersonReID/ResNet50_strong_baseline_market1501.yaml
@ -1,177 +0,0 @@
-# global configs
-Global:
-  checkpoints: null
-  pretrained_model: null
-  output_dir: "./output/"
-  device: "gpu"
-  save_interval: 40
-  eval_during_train: True
-  eval_interval: 10
-  epochs: 120
-  print_batch_step: 20
-  use_visualdl: False
-  warmup_by_epoch: True
-  eval_mode: "retrieval"
-  re_ranking: True
-  # used for static mode and model export
-  image_shape: [3, 256, 128]
-  save_inference_dir: "./inference"
-
-# model architecture
-Arch:
-  name: "RecModel"
-  infer_output_key: "features"
-  infer_add_softmax: False
-  Backbone:
-    name: "ResNet50_last_stage_stride1"
-    pretrained: True
-    stem_act: null
-  BackboneStopLayer:
-    name: "flatten"
-  Neck:
-    name: BNNeck
-    num_features: &feat_dim 2048
-  Head:
-    name: "FC"
-    embedding_size: *feat_dim
-    class_num: &class_num 751
-    weight_attr:
-      initializer:
-        name: Normal
-        std: 0.001
-    bias_attr: False
-
-# loss function config for traing/eval process
-Loss:
-  Train:
-    - CELoss:
-        weight: 1.0
-        epsilon: 0.1
-    - TripletLossV3:
-        weight: 1.0
-        margin: 0.3
-        normalize_feature: false
-    - CenterLoss:
-        weight: 0.0005
-        num_classes: *class_num
-        feat_dim: *feat_dim
-  Eval:
-    - CELoss:
-        weight: 1.0
-
-Optimizer:
-  - Adam:
-      scope: model
-      lr:
-        name: Piecewise
-        decay_epochs: [30, 60]
-        values: [0.00035, 0.000035, 0.0000035]
-        warmup_epoch: 10
-        warmup_start_lr: 0.0000035
-        warmup_by_epoch: True
-      regularizer:
-        name: 'L2'
-        coeff: 0.0005
-  - SGD:
-      scope: CenterLoss
-      lr:
-        name: Constant
-        learning_rate: 1000.0
-
-# data loader for train and eval
-DataLoader:
-  Train:
-    dataset:
-        name: "Market1501"
-        image_root: "./dataset/Market-1501-v15.09.15"
-        cls_label_path: "bounding_box_train"
-        transform_ops:
-          - DecodeImage:
-              to_rgb: True
-              channel_first: False
-          - ResizeImage:
-              size: [128, 256]
-          - RandFlipImage:
-              flip_code: 1
-          - Pad:
-              padding: 10
-          - RandCropImage:
-              size: [128, 256]
-              scale: [ 0.8022, 0.8022 ]
-              ratio: [ 0.5, 0.5 ]
-          - NormalizeImage:
-              scale: 0.00392157
-              mean: [0.485, 0.456, 0.406]
-              std: [0.229, 0.224, 0.225]
-              order: ''
-          - RandomErasing:
-              EPSILON: 0.5
-              sl: 0.02
-              sh: 0.4
-              r1: 0.3
-              mean: [0.4914, 0.4822, 0.4465]
-    sampler:
-        name: DistributedRandomIdentitySampler
-        batch_size: 64
-        num_instances: 4
-        drop_last: True
-        shuffle: True
-    loader:
-        num_workers: 4
-        use_shared_memory: True
-  Eval:
-    Query:
-      dataset:
-        name: "Market1501"
-        image_root: "./dataset/Market-1501-v15.09.15"
-        cls_label_path: "query"
-        transform_ops:
-          - DecodeImage:
-              to_rgb: True
-              channel_first: False
-          - ResizeImage:
-              size: [128, 256]
-          - NormalizeImage:
-              scale: 0.00392157
-              mean: [0.485, 0.456, 0.406]
-              std: [0.229, 0.224, 0.225]
-              order: ''
-      sampler:
-        name: DistributedBatchSampler
-        batch_size: 128
-        drop_last: False
-        shuffle: False
-      loader:
-        num_workers: 4
-        use_shared_memory: True
-
-    Gallery:
-      dataset:
-        name: "Market1501"
-        image_root: "./dataset/Market-1501-v15.09.15"
-        cls_label_path: "bounding_box_test"
-        transform_ops:
-          - DecodeImage:
-              to_rgb: True
-              channel_first: False
-          - ResizeImage:
-              size: [128, 256]
-          - NormalizeImage:
-              scale: 0.00392157
-              mean: [0.485, 0.456, 0.406]
-              std: [0.229, 0.224, 0.225]
-              order: ''
-      sampler:
-        name: DistributedBatchSampler
-        batch_size: 128
-        drop_last: False
-        shuffle: False
-      loader:
-        num_workers: 4
-        use_shared_memory: True
-
-Metric:
-  Eval:
-    - Recallk:
-        topk: [1, 5]
-    - mAP: {}
--- a/ppcls/engine/evaluation/retrieval.py
+++ b/ppcls/engine/evaluation/retrieval.py
@ -82,6 +82,7 @@ def retrieval_eval(engine, epoch_id=0):
                        metric_dict[key] += metric_tmp[key] * block_fea.shape[
                            0] / len(query_feas)
        else:
+            metric_dict = dict()
            distmat = re_ranking(
                query_feas, gallery_feas, k1=20, k2=6, lambda_value=0.3)
            cmc, mAP = eval_func(distmat,
@ -93,9 +94,6 @@ def retrieval_eval(engine, epoch_id=0):
            metric_dict["recall5(RK)"] = cmc[4]
            metric_dict["mAP(RK)"] = mAP

-            for key in metric_tmp:
-                metric_dict[key] = metric_tmp[key] * block_fea.shape[0] / len(
-                    query_feas)
    metric_info_list = []
    for key in metric_dict:
        if metric_key is None: