If there is one training image then will the network be trained on one training image which is modified by TLT or be trained on two training images , one original, one from augmentation ? Or will there be one original and two from augmentation ? If it’s trained only on one image , which augmentation was done - I have listed the augmentation I used below?
Below is my augmentation from the config file:
augmentation_config {
preprocessing {
output_image_width: 840
output_image_height: 480
min_bbox_width: 1.0
min_bbox_height: 1.0
output_image_channel: 3
}
spatial_augmentation {
hflip_probability: 0.5
zoom_min: 1.0
zoom_max: 1.0
translate_max_x: 8.0
translate_max_y: 8.0
}
color_augmentation {
hue_rotation_max: 25.0
saturation_shift_max: 0.20000000298
contrast_scale_max: 0.10000000149
contrast_center: 0.5
}
}