Skip to content
GitLab
Select Git revision
  • 3.x
  • 4.x
  • remplacer_assetic_par_webpack
  • 5-migration-vers-la-nouvelle-architecture-basee-sur-webpack-npm
  • 8-enrichir-le-dossier-test-sur-la-base-de-la-nouvelle-architecture
  • 6-supprimer-isdevtools
  • 8-enrichir-le-dossier-test-sur-la-base-de-la-nouvelle-architecture-2
  • master default
  • symfony3.4_ramsey_compatibility
  • 1.x
  • 0.1.x
  • 2.1.0
  • 2.0.2
  • 2.0.1
  • 2.0.0
  • 1.0.1
  • 0.1.5
  • 1.0.0
  • 0.1.4
  • 0.1.3
  • 0.1.2
  • 0.1.1
  • 0.1
Find file BlameHistoryPermalink
CREATEYOUROWN.md 2.61 KiB

Create your own architecture

This section gives a few tips to create your own models ready to be used in inference using OTBTF's TensorflowModelServe and TensorflowModelTrain applications.

Model inputs

Dimensions

All networks must input 4D tensors.

  • dim 0 is for the batch dimension. It is used in the TensorflowModelTrain application during training, and in patch-based mode during inference: in this mode, TensorflowModelServe performs the inference of several patches simultaneously. In fully-convolutional mode, a single slice of the batch dimension is used.
  • dim 1 and 2 are for the spatial dimensions,
  • dim 3 is for the image channels. Even if your image have only 1 channel, you must set a shape value equals to 1 for the last dimension of the input placeholder.

Shapes

For nets intended to work in patch-based mode, you can stick with a placeholder where you define your patch size explicitly in dim 1 and dim 2. However, for nets intended to work in fully-convolutional mode, you must set None in dim 1 and dim 2 (before Tensorflow 2.X, it was possible to feed placeholders with a tensor of different size where the dims were defined, but no more after!). For instance, let consider an input raster with 4 spectral bands: the input shape of the model input would be like [None, None, None, 4] to work in fully-convolutional mode. By doing so, the use of input images of any size is enabled (TensorflowModelServe will automatically compute the input/output regions sizes to process, given the receptive field and expression field of your net).

Model outputs

Dimensions

Supported tensors for the outputs must have between 2 and 4 dimensions. OTBTF always consider that the size of the last dimension is the number of channels in the output. For instance, you can have a model that outputs 8 channels with a tensor of shape [None, 8] or [None, None, None, 8]

Name your tensors and nodes

Always name explicitly your models outputs. You will need the output tensor name for performing the inference with TensoflowModelServe. If you forget to name them, use the graph viewer in tensorboard to get the names.

Training

If you want to enable your network training with the TensorflowModelTrain application, do not forget to name your optimizers/operators! You can build a single operator from multiple ones using the tf.group command, which also enable you to name your new operator. For sequential nodes trigger, you can build an operator that do what you want is the desired order using the tf.control_dependancies with TF <= 1.15.