:orphan:

#################################
Level 12: Optimize training speed
#################################

In this level you'll use compilers, advanced profilers and mixed precision techniques to train bigger models faster.

----

.. raw:: html

    <div class="display-card-container">
        <div class="row">

.. displayitem::
   :header: Speed up models by compiling them
   :description:  Use torch.compile to speed up models on modern hardware
   :col_css: col-md-4
   :button_link: ../advanced/compile.html
   :height: 150
   :tag: intermediate

.. displayitem::
   :header: Explore advanced mixed precision settings
   :description: Enable state-of-the-art scaling with advanced mix-precision settings.
   :col_css: col-md-4
   :button_link: ../common/precision_intermediate.html
   :height: 150
   :tag: intermediate

.. displayitem::
   :header: Enable advanced profilers
   :description: Tune model performance with profilers.
   :col_css: col-md-4
   :button_link: ../tuning/profiler_basic.html
   :height: 150
   :tag: intermediate

.. displayitem::
   :header: Profile PyTorch operations
   :description: Learn to find bottlenecks in PyTorch operations.
   :col_css: col-md-4
   :button_link: ../tuning/profiler_intermediate.html
   :height: 150
   :tag: intermediate

.. raw:: html

        </div>
    </div>