为什么来自 Tensorflow 2 对象检测 API 的微调模型的低 mAP？

Question

我遵循了所有步骤并在线阅读了所有内容，并且我从 TF2 OD 的模型动物园成功训练了 SSD-MobileNetV1 API。

我用新的classes“手枪”和“刀”微调了这个模型，我使用了一个包含 3500 张图像的平衡数据集。培训进行得很好，但是当我运行使用“pascal_voc_detection_metrics”进行评估过程（用于验证）时，我达到了 0.005 AP@0.5（检测模型仅达到或多或少的 0.005 AP） class“手枪”非常低，但 0.93 AP@0.5 与 class“刀”。

我不明白为什么。我真的阅读了所有内容，但找不到解决方案。

SDD-MobileNetV1 的配置：

model {
  ssd {
    num_classes: 2
    image_resizer {
      fixed_shape_resizer {
        height: 640
        width: 640
      }
    }
    feature_extractor {
      type: "ssd_mobilenet_v1_fpn_keras"
      depth_multiplier: 1.0
      min_depth: 16
      conv_hyperparams {
        regularizer {
          l2_regularizer {
            weight: 4e-05
          }
        }
        initializer {
          random_normal_initializer {
            mean: 0.0
            stddev: 0.01
          }
        }
        activation: RELU_6
        batch_norm {
          decay: 0.997
          scale: true
          epsilon: 0.001
        }
      }
      override_base_feature_extractor_hyperparams: true
      fpn {
        min_level: 3
        max_level: 7
      }
    }
    box_coder {
      faster_rcnn_box_coder {
        y_scale: 10.0
        x_scale: 10.0
        height_scale: 5.0
        width_scale: 5.0
      }
    }
    matcher {
      argmax_matcher {
        matched_threshold: 0.5
        unmatched_threshold: 0.5
        ignore_thresholds: false
        negatives_lower_than_unmatched: true
        force_match_for_each_row: true
        use_matmul_gather: true
      }
    }
    similarity_calculator {
      iou_similarity {
      }
    }
    box_predictor {
      weight_shared_convolutional_box_predictor {
        conv_hyperparams {
          regularizer {
            l2_regularizer {
              weight: 4e-05
            }
          }
          initializer {
            random_normal_initializer {
              mean: 0.0
              stddev: 0.01
            }
          }
          activation: RELU_6
          batch_norm {
            decay: 0.997
            scale: true
            epsilon: 0.001
          }
        }
        depth: 256
        num_layers_before_predictor: 4
        kernel_size: 3
        class_prediction_bias_init: -4.6
      }
    }
    anchor_generator {
      multiscale_anchor_generator {
        min_level: 3
        max_level: 7
        anchor_scale: 4.0
        aspect_ratios: 1.0
        aspect_ratios: 2.0
        aspect_ratios: 0.5
        scales_per_octave: 2
      }
    }
    post_processing {
      batch_non_max_suppression {
        score_threshold: 1e-08
        iou_threshold: 0.6
        max_detections_per_class: 100
        max_total_detections: 100
        use_static_shapes: false
      }
      score_converter: SIGMOID
    }
    normalize_loss_by_num_matches: true
    loss {
      localization_loss {
        weighted_smooth_l1 {
        }
      }
      classification_loss {
        weighted_sigmoid_focal {
          gamma: 2.0
          alpha: 0.25
        }
      }
      classification_weight: 1.0
      localization_weight: 1.0
    }
    encode_background_as_zeros: true
    normalize_loc_loss_by_codesize: true
    inplace_batchnorm_update: true
    freeze_batchnorm: false
  }
}
train_config {
  batch_size: 4
  data_augmentation_options {
    random_horizontal_flip {
    }
  }
  data_augmentation_options {
    random_crop_image {
      min_object_covered: 0.0
      min_aspect_ratio: 0.75
      max_aspect_ratio: 3.0
      min_area: 0.75
      max_area: 1.0
      overlap_thresh: 0.0
    }
  }
  sync_replicas: true
  optimizer {
    momentum_optimizer {
      learning_rate {
        cosine_decay_learning_rate {
          learning_rate_base: 0.04
          total_steps: 25000
          warmup_learning_rate: 0.013333
          warmup_steps: 2000
        }
      }
      momentum_optimizer_value: 0.9
    }
    use_moving_average: false
  }
  fine_tune_checkpoint: "pre-trained-models/ssd_mobilenet_v1_fpn_640x640_coco17_tpu-8/checkpoint/ckpt-0"
  num_steps: 25000
  startup_delay_steps: 0.0
  replicas_to_aggregate: 8
  max_number_of_boxes: 100
  unpad_groundtruth_tensors: false
  fine_tune_checkpoint_type: "detection"
  fine_tune_checkpoint_version: V2
}
train_input_reader {
  label_map_path: "/annotations/label_map.pbtxt"
  tf_record_input_reader {
    input_path: "/annotations/train.record"
  }
}
eval_config {
  metrics_set: "pascal_voc_detection_metrics"
  use_moving_averages: false
  batch_size: 1
}
eval_input_reader {
  label_map_path: "/annotations/label_map.pbtxt"
  shuffle: false
  num_epochs: 1
  tf_record_input_reader {
    input_path: "/annotations/validation.record"
  }
}

我使用 model_main_tf2.py 进行训练和评估，并使用 roboflow 在 TFRecords 中转换我的图像。

Answer 1

这是 link 上报告的库错误。 COCO 指标没有这个问题，所以用它来评估你的模型。问题还没有解决。如果您想关注对代码所做的更新（它们工作正常），请关注之前的 link 以及这个 link

为什么来自 Tensorflow 2 对象检测 API 的微调模型的低 mAP？

Why low mAP on fine-tuned model from Tensorflow 2 Object Detection API?

python

object-detection

tensorflow

object-detection-api