index.rst

OneFlow API Reference
===================================


Distributed performance (high efficiency) is the core technical difficulty of deep learning frameworks. 

OneFlow upholds the core concept and architecture of static compilation and streaming parallelism around performance improvement and heterogeneous distributed scaling, solving the challenge of memory wall at cluster level with world-leading technology.


.. toctree::
    :maxdepth: 1

    troubleshooting



.. toctree::
    :maxdepth: 1
    :caption: OneFlow Python API

    oneflow
    nn
    nn.functional
    tensor
    tensor_attributes
    type_info
    autograd
    cuda
    distributed
    distributions
    hub
    linalg
    nn.init
    optim
    graph
    auto_parallel
    image
    utils.data
    utils.global_view
    utils.tensor
    one_embedding
    environment_variables



Indices and tables
==================

* :ref:`genindex`
* :ref:`modindex`
* :ref:`search`