MXNet API for R

This section contains documentation for the R version of MXNet.

To instead view documentation directly in your R console, enter:

library(mxnet); help(mxnet)

and then click on Index in the Help window that appears.

• NDArray
  • The core MXNet data structure for all mathematical operations
  • Get array attributes
  • Create arrays
  • Manipulation of arrays
  • Mathematical operations on arrays
  • Neural network array operations
• Symbol
  • Symbolic programming with computation graphs
  • Get symbol attributes
  • Create symbols
  • Manipulation of symbols
  • Mathematical operations on symbols
  • Neural network symbol operations
• IO
  • Efficient distributed data loading and augmentation
  • Loading image data
• Context
  • Manage device type and id where computations are carried out
• Model
  • Create, load/save, and use MXNet models
• Metric
  • Evaluation metrics to measure model performance
• Optimization
  • Initialize and update model weights during training
  • Optimizers
  • Initialization
  • Learning rate schedule
  • Optimizer updates (NDArray)
  • Optimizer updates (Symbol)
• Random Sampling
  • Random number generation in MXNet
  • Random NDArrays
  • Random Symbols
• Monitoring
  • Callback: Functions to track various status during an epoch
  • Profiler: Running time and memory consumption of MXNet models
• Advanced
  • Key-Value Store: Operate over multiple devices (GPUs) on a single device
  • Executor: Internal classes for managing symbolic graph execution
• Complete Library
  • List of all MXNet functions available in R