Metadata-Version: 2.1
Name: tetra-hub
Version: 0.3.2
Summary: Python API for Tetra Hub.
Home-page: https://tetra.ai/
Author: Tetra Intelligence Systems Inc.
Author-email: support@tetra.ai
License: UNKNOWN
Platform: UNKNOWN
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: Operating System :: MacOS :: MacOS X
Classifier: Operating System :: POSIX :: Linux
Classifier: Operating System :: Microsoft :: Windows
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Topic :: Scientific/Engineering
Classifier: Topic :: Software Development
Requires-Python: >=3.7
Requires-Dist: requests
Requires-Dist: deprecation
Requires-Dist: requests-toolbelt
Requires-Dist: tqdm
Requires-Dist: protobuf (<4,>=3.15)
Requires-Dist: numpy (<2,>=1.22.0)
Requires-Dist: h5py (<4,>=3.6.0)
Provides-Extra: coremltools
Requires-Dist: coremltools (==5.2.0) ; extra == 'coremltools'
Provides-Extra: torch
Requires-Dist: torch (==1.11.0) ; (sys_platform == "darwin") and extra == 'torch'
Requires-Dist: torchvision (==0.12) ; (sys_platform == "darwin") and extra == 'torch'
Requires-Dist: torch (==1.12.0) ; (sys_platform == "linux") and extra == 'torch'
Requires-Dist: torchvision (==0.13) ; (sys_platform == "linux") and extra == 'torch'

Tetra Hub
=========

`Tetra Hub <https://tetra.ai>`_ provides transparency into the
on-device performance of machine learning models. It allows quick examination
into key metrics to improve the model.


Tetra Hub allows users to analyze a model throught the following steps:

- Simply upload a model and submit a profiling job, targeting a specific
  device.
- The model is compiled and tailored to the target platform.
- A physical device is provisioned in the cloud. Tetra Hub analysis tools run
  the model on the device and gather metrics such as which model layers run
  on which compute units, model latency, and peak memory.
- The results are displayed on Tetra Hub, providing insights to understand
  the performance and opportunities for further improvements.

:code:`tetra_hub` is a python package that provides an API for users to upload a model,
submit the profile jobs for hardware and get key metrics to optimize the
machine learning model further.


Installation with PyPI
----------------------

The easiest way to install :code:`tetra_hub` is by using pip, running

:code:`pip install tetra-hub`

For more information, check out the `documentation <https://hub.tetra.ai/docs/>`_ for more information on `tetra_hub` APIs, how to use them, and for examples.

License
-------

Copyright (c) 2022, Tetra Intelligence Systems Inc. All rights reserved.


