Welcome to PADL’s documentation!

Pipeline abstractions for deep learning.

---

PADL:

• is a pipeline builder for PyTorch.

• can be used with all of the great PyTorch functionality you’re used to for writing layers.

• allows users to build pre-processing, forward passes, loss functions and post-processing into the pipeline.

• models may have arbitrary topologies and make use of arbitrary packages from the python ecosystem.

• allows for converting standard functions to PADL components using a single keyword transform.

PADL was developed at LF1, an AI innovation lab based in Berlin, Germany.

Contents:

• Introduction
  • Getting Started
  • Contributing
  • Licensing
• Usage
  • Transforms and Pipelines
  • Creating Transforms
  • Combining Transforms into Pipelines
  • Stages: Preprocess, Forward and Postprocess
  • Applying Transforms to Data
  • Saving and Loading
  • Printing and Slicing
  • Using PyTorch Modules with Transforms
  • Debugging your Transforms
  • Extras
  • Import anything you need from padl
• Advanced Topics
  • Saving
• API Documentation
  • padl
  • padl.dumptools
  • padl.dumptools.ast_utils
  • padl.dumptools.inspector
  • padl.dumptools.packagefinder
  • padl.dumptools.serialize
  • padl.dumptools.sourceget
  • padl.dumptools.symfinder
  • padl.dumptools.var2mod
  • padl.exceptions
  • padl.print_utils
  • padl.transforms
  • padl.util_transforms
  • padl.utils
  • padl.version
  • padl.wrap