Module: tfp.experimental.auto_batching.frontend

View source on GitHub

AutoGraph-based auto-batching frontend.

Modules

ab_type_inference module: Type inference pass on functional control flow graph.

allocation_strategy module: Live variable analysis.

dsl module: Python-embedded DSL frontend for authoring autobatchable IR programs.

instructions module: Instruction language for auto-batching virtual machine.

lowering module: Lowering the full IR to stack machine instructions.

st module: A stackless auto-batching VM.

stack module: Optimizing stack usage (pushes and pops).

tf_backend module: TensorFlow (graph) backend for auto-batching VM.

vm module: The auto-batching VM itself.

Classes

class Context: Context object for auto-batching multiple Python functions together.

Functions

truthy(...): Normalizes Tensor ranks for use in if conditions.

Other Members

  • TF_BACKEND