TensorFlow目标检测训练出现-1 mAP、检测结果为N/A问题求助

在Colab上遵循目标检测API教程训练自定义目标检测模型，采用SSD MobileNet V2架构，使用2083张训练样本、263张测试样本。训练完成后通过TensorBoard查看，模型似乎能检测到部分物体，但评估时仅得到**-1 mAP**，且检测结果全为"N/A"、置信度100%（见下图）。

以下是模型配置文件（完整相关文件见我的GitHub仓库）：

model {
  ssd {
    num_classes: 8
    image_resizer {
      fixed_shape_resizer {
        height: 300
        width: 300
      }
    }
    feature_extractor {
      type: "ssd_mobilenet_v2_keras"
      depth_multiplier: 1.0
      min_depth: 16
      conv_hyperparams {
        regularizer {
          l2_regularizer {
            weight: 3.9999998989515007e-05
          }
        }
        initializer {
          truncated_normal_initializer {
            mean: 0.0
            stddev: 0.029999999329447746
          }
        }
        activation: RELU_6
        batch_norm {
          decay: 0.9700000286102295
          center: true
          scale: true
          epsilon: 0.0010000000474974513
          train: true
        }
      }
      override_base_feature_extractor_hyperparams: true
    }
    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 {
      convolutional_box_predictor {
        conv_hyperparams {
          regularizer {
            l2_regularizer {
              weight: 3.9999998989515007e-05
            }
          }
          initializer {
            random_normal_initializer {
              mean: 0.0
              stddev: 0.009999999776482582
            }
          }
          activation: RELU_6
          batch_norm {
            decay: 0.9700000286102295
            center: true
            scale: true
            epsilon: 0.0010000000474974513
            train: true
          }
        }
        min_depth: 0
        max_depth: 0
        num_layers_before_predictor: 0
        use_dropout: false
        dropout_keep_probability: 0.800000011920929
        kernel_size: 1
        box_code_size: 4
        apply_sigmoid_to_scores: false
        class_prediction_bias_init: -4.599999904632568
      }
    }
    anchor_generator {
      ssd_anchor_generator {
        num_layers: 6
        min_scale: 0.20000000298023224
        max_scale: 0.949999988079071
        aspect_ratios: 1.0
        aspect_ratios: 2.0
        aspect_ratios: 0.5
        aspect_ratios: 3.0
        aspect_ratios: 0.33329999446868896
      }
    }
    post_processing {
      batch_non_max_suppression {
        score_threshold: 9.99999993922529e-09
        iou_threshold: 0.6000000238418579
        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 {
          delta: 1.0
        }
      }
      classification_loss {
        weighted_sigmoid_focal {
          gamma: 2.0
          alpha: 0.75
        }
      }
      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 {
    ssd_random_crop {
    }
  }
  sync_replicas: true
  optimizer {
    momentum_optimizer {
      learning_rate {
        cosine_decay_learning_rate {
          learning_rate_base: 0.001
          total_steps: 25000
          warmup_learning_rate: 0.0001
          warmup_steps: 2500
        }
      }
      momentum_optimizer_value: 0.8999999761581421
    }
    use_moving_average: false
  }
  fine_tune_checkpoint: "pre-trained-models/ssd_mobilenet_v2_320x320_coco17_tpu-8/checkpoint/ckpt-0"
  num_steps: 25000
  startup_delay_steps: 0.0
  replicas_to_aggregate: 8
  max_number_of_boxes: 10
  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: "coco_detection_metrics"
  use_moving_averages: false
  batch_size: 1
  eval_interval_secs: 30
  num_examples: 236 # no of test images
  num_visualizations: 10 # no of visualizations for tensorboard
  max_num_boxes_to_visualize: 5
  visualize_groundtruth_boxes: true
}
eval_input_reader {
  label_map_path: "annotations/label_map.pbtxt"
  shuffle: true
  num_epochs: 1
  tf_record_input_reader {
    input_path: "annotations/test_*.record"
  }
}

已尝试的排查方法：

• 修改TFRecord文件生成代码
• 调整batch size和训练步数
• 将label_map.pbtxt内容改为全拉丁字符
• 检查TFRecord文件是否正常生成

问题未解决，且训练至100~200步时，评估就出现-1 mAP、检测结果为"N/A"的情况。

内容的提问来源于stack exchange，提问作者muheonkom