What does mean «train_config» → «data_augmentation_options» in TensorFlow?

dmitrii_fediuk · May 18, 2019, 1:38pm

tensorflow/models/blob/v1.13.0/research/object_detection/samples/configs/faster_rcnn_resnet101_voc07.config#L109-L112


data_augmentation_options {
  random_horizontal_flip {
  }
}

dmitrii_fediuk · May 18, 2019, 1:42pm

The data_augmentation_options in train_config can be used to specify how training data can be modified. This field is optional.

github.com/tensorflow/models/blob/v1.13.0/research/object_detection/g3doc/configuring_jobs.md#input-preprocessing

The list of options:

github.com

tensorflow/models/blob/v1.13.0/research/object_detection/protos/preprocessor.proto#L5-L38


      
          // Message for defining a preprocessing operation on input data.
          // See: //third_party/tensorflow_models/object_detection/core/preprocessor.py
          message PreprocessingStep {
            oneof preprocessing_step {
              NormalizeImage normalize_image = 1;
              RandomHorizontalFlip random_horizontal_flip = 2;
              RandomPixelValueScale random_pixel_value_scale = 3;
              RandomImageScale random_image_scale = 4;
              RandomRGBtoGray random_rgb_to_gray = 5;
              RandomAdjustBrightness random_adjust_brightness = 6;
              RandomAdjustContrast random_adjust_contrast = 7;
              RandomAdjustHue random_adjust_hue = 8;
              RandomAdjustSaturation random_adjust_saturation = 9;
              RandomDistortColor random_distort_color = 10;
              RandomJitterBoxes random_jitter_boxes = 11;
              RandomCropImage random_crop_image = 12;
              RandomPadImage random_pad_image = 13;
              RandomCropPadImage random_crop_pad_image = 14;
              RandomCropToAspectRatio random_crop_to_aspect_ratio = 15;
              RandomBlackPatches random_black_patches = 16;

This file has been truncated. show original

stackoverflow.com/a/46901051