#!/usr/bin/python3

  # Simple while loop
  a = 0
  while a < 15:
      print(a, end=' ')
      if a == 10:
          print("made it to ten!!")
      a = a + 1
  print()

Hybrid Front-End

A new hybrid front-end provides ease-of-use and flexibility in eager mode, while seamlessly transitioning to graph mode for speed, optimization, and functionality in C++ runtime environments.

Distributed Training

Take advantage of native support for asynchronous execution of collective operations and peer-to-peer communication that is accessible from both Python and C++.

  #!/usr/bin/python3

  # Simple while loop
  a = 0
  while a < 15:
      print(a, end=' ')
      if a == 10:
          print("made it to ten!!")
      a = a + 1
  print()

  #!/usr/bin/python3

  # Simple while loop
  a = 0
  while a < 15:
      print(a, end=' ')
      if a == 10:
          print("made it to ten!!")
      a = a + 1
  print()

Python-First

Deep integration into Python allows popular libraries and packages to be used, while a new pure C++ interface (beta) enables performance-critical research.

Tools & Libraries

Access a rich ecosystem of tools and libraries to extend PyTorch and support development in areas from computer vision to reinforcement learning.

  #!/usr/bin/python3

  # Simple while loop
  a = 0
  while a < 15:
      print(a, end=' ')
      if a == 10:
          print("made it to ten!!")
      a = a + 1
  print()

  #!/usr/bin/python3

  # Simple while loop
  a = 0
  while a < 15:
      print(a, end=' ')
      if a == 10:
          print("made it to ten!!")
      a = a + 1
  print()

Native ONNX Support

Export models in the standard ONNX (Open Neural Network Exchange) format for direct access to ONNX-compatible platforms, runtimes, visualizers, and more.

Cloud Partners

Get up and running quickly with PyTorch through cloud platforms for training and inference.

  #!/usr/bin/python3

  # Simple while loop
  a = 0
  while a < 15:
      print(a, end=' ')
      if a == 10:
          print("made it to ten!!")
      a = a + 1
  print()