Changes for version 0.0.7 - 2023-10-05

  • Features
    • Add object detection demo. See <https://github.com/EntropyOrg/perl-AI-TensorFlow-Libtensorflow/pull/23>.
  • Refactoring
    • Add timer to the notebooks to time the inference steps. See <https://github.com/EntropyOrg/perl-AI-TensorFlow-Libtensorflow/pull/17>.
  • Documentation
    • Add information about installing GPU version of `libtensorflow` either on the "bare metal" or with Docker GPU runtime support. See <https://github.com/EntropyOrg/perl-AI-TensorFlow-Libtensorflow/pull/18>.
  • Build changes
    • Add Dockerfile that builds GPU version of the omnibus notebook image. Update the CI to additionally build the GPU Docker image. See <https://github.com/EntropyOrg/perl-AI-TensorFlow-Libtensorflow/pull/16>.

Documentation

List of functions exported by TensorFlow C API
GPU-specific installation and usage information.
Using TensorFlow to do object detection using a pre-trained model
Using TensorFlow to do gene expression prediction using a pre-trained model
Using TensorFlow to do image classification using a pre-trained model
Start here for an overview of the library

Modules

Bindings for Libtensorflow deep learning library
Maps Operation to API description
Buffer that holds pointer to data with length
A list of devices available for the session to run on
A TensorFlow computation, represented as a dataflow graph
Holds options that can be passed to ::Graph::ImportGraphDef
Results from importing a graph definition
Input of operation as (operation, index) pair
Private class for AI::TensorFlow::Libtensorflow
Type to hold string list as void** strings, size_t* lengths, int num_items
Type to hold pointer and size in a scalar (input only)
Type to hold pointer and size in a scalar reference
ArrayRef to pack()'ed scalar argument with size argument (as int)
Maybe[ArrayRef] to pack()'ed scalar argument with size argument (as int) (size is -1 if undef)
Turn FFI::Platypus::Record into packed array (+ size)?
[private] Allocation utilities
Output of operation as (operation, index) pair
Session for driving ::Graph execution
Holds options that can be passed during session creation
Status used for error checking
TensorFlow dynamic library handle and ops
A variable-capacity string type
A multi-dimensional array of elements of a single data type