TensorFlow Lite converter

TensorFlow Lite uses the optimized FlatBuffer format to represent graphs. Therefore, a TensorFlow model (protocol buffer) needs to be converted into a FlatBuffer file before deploying to clients.

From model training to device deployment

The TensorFlow Lite converter generates a TensorFlow Lite FlatBuffer file (.tflite) from a TensorFlow model.

The converter supports the following input formats:

  • SavedModels
  • Frozen GraphDef: Models generated by freeze_graph.py.
  • tf.keras HDF5 models.
  • Any model taken from a tf.Session (Python API only).

The TensorFlow Lite FlatBuffer file is then deployed to a client device (generally a mobile or embedded device), and the TensorFlow Lite interpreter uses the compressed model for on-device inference. This conversion process is shown in the diagram below:

TFLite converter workflow


The TensorFlow Lite Converter can be used from either of these two options:

  • Python (Preferred): Using the Python API makes it easier to convert models as part of a model development pipeline, and helps mitigate compatibility issues early on.
  • Command line