Question

我正在尝试使用tensorflow api预训练模型创建自己的对象检测器： faster_rcnn_inception_v2_coco ，并且我正在与Google-Colab合作。

我做的第一件事是收集培训数据，并测试数据以进行验证。训练文件夹包含336个xml，而测试文件夹包含72个。

注意：我自己是通过翻转数据来对训练数据进行数据增强的，因此从技术上讲，在训练中有168张图像实际上已翻倍。

将其转换为唱片之后，我就开始用model_main.py训练模型：

!python object_detection/model_main.py --model_dir=training --pipeline_config_path=training/faster_rcnn_inception_v2_coco.config

我的配置文件如下：

# Faster R-CNN with Inception v2, configuration for MSCOCO Dataset.
# Users should configure the fine_tune_checkpoint field in the train config as
# well as the label_map_path and input_path fields in the train_input_reader and
# eval_input_reader. Search for "PATH_TO_BE_CONFIGURED" to find the fields that
# should be configured.


model {
  faster_rcnn {
    num_classes: 2
    image_resizer {
      keep_aspect_ratio_resizer {
        min_dimension: 600
        max_dimension: 1024
      }
    }
    feature_extractor {
      type: 'faster_rcnn_inception_v2'
      first_stage_features_stride: 16
    }
    first_stage_anchor_generator {
      grid_anchor_generator {
        scales: [0.25, 0.5, 1.0, 2.0]
        aspect_ratios: [0.5, 1.0, 2.0]
        height_stride: 16
        width_stride: 16
      }
    }
    first_stage_box_predictor_conv_hyperparams {
      op: CONV
      regularizer {
        l2_regularizer {
          weight: 0.0
        }
      }
      initializer {
        truncated_normal_initializer {
          stddev: 0.01
        }
      }
    }
    first_stage_nms_score_threshold: 0.0
    first_stage_nms_iou_threshold: 0.7
    first_stage_max_proposals: 300
    first_stage_localization_loss_weight: 2.0
    first_stage_objectness_loss_weight: 1.0
    initial_crop_size: 14
    maxpool_kernel_size: 2
    maxpool_stride: 2
    second_stage_box_predictor {
      mask_rcnn_box_predictor {
        use_dropout: false
        dropout_keep_probability: 1.0
        fc_hyperparams {
          op: FC
          regularizer {
            l2_regularizer {
              weight: 0.0
            }
          }
          initializer {
            variance_scaling_initializer {
              factor: 1.0
              uniform: true
              mode: FAN_AVG
            }
          }
        }
      }
    }
    second_stage_post_processing {
      batch_non_max_suppression {
        score_threshold: 0.0
        iou_threshold: 0.6
        max_detections_per_class: 100
        max_total_detections: 300
      }
      score_converter: SOFTMAX
    }
    second_stage_localization_loss_weight: 2.0
    second_stage_classification_loss_weight: 1.0
  }
}

train_config: {
  batch_size: 1
  optimizer {
    momentum_optimizer: {
      learning_rate: {
        manual_step_learning_rate {
          initial_learning_rate: 0.0002
          schedule {
            step: 900000
            learning_rate: .00002
          }
          schedule {
            step: 1200000
            learning_rate: .000002
          }
        }
      }
      momentum_optimizer_value: 0.9
    }
    use_moving_average: false
  }
  gradient_clipping_by_norm: 10.0
  fine_tune_checkpoint: "faster_rcnn_inception_v2_coco_2018_01_28/model.ckpt"
  fine_tune_checkpoint_type: "detection"
  from_detection_checkpoint: true
  # Note: The below line limits the training process to 200K steps, which we
  # empirically found to be sufficient enough to train the COCO dataset. This
  # effectively bypasses the learning rate schedule (the learning rate will
  # never decay). Remove the below line to train indefinitely.
  num_steps: 200000
  data_augmentation_options {
    random_horizontal_flip {
    }
  }
}

train_input_reader: {
  tf_record_input_reader {
    input_path: "data/train.record"
  }
  label_map_path: "training/object-detection.pbtxt"
}

eval_config: {
  num_examples: 72
  # Note: The below line limits the evaluation process to 10 evaluations.
  # Remove the below line to evaluate indefinitely.
   max_evals: 10
}

eval_input_reader: {
  tf_record_input_reader {
    input_path: "data/test.record"
  }
  label_map_path: "training/object-detection.pbtxt"
  shuffle: false
  num_readers: 1
}

当我停止训练时，我打开了Tensorboard并得到了以下图表：

Training Loss

Validation Loss

据我所知，当过度拟合训练数据时，训练损失应该从高开始，然后逐渐降低，而验证损失应该从高开始，然后下降，然后再次上升。

但是我不确定为什么我的验证损失会从低开始，而只会上升，所以知道为什么吗？另外，我看到我在此图形上有少量图形点，也许会影响图形？

我必须说，当我用随机的新图像填充模型以预测结果非常好时，这对我来说很奇怪。

Tensorflow对象检测API-验证丢失率从低到高

0 个答案: