NAS + RL
Project description
<br> <br> <br> <br>
<img align=”center” width=”350” height=”200” src=”https://i.ibb.co/4Zf5mvG/logo.png” alt=”convstruct-github-logo” border=”0”>
<br> <br> <br> <br>
—
[<img align=”center” src=”https://i.ibb.co/zmFZZDW/hero-banner.png” alt=”convstruct-hero-banner” border=”0”>](###Convstruct-roadmap)
—
<br>
###Convstruct 1.1 is here :zap: :zap:
1.1 is an architecture update as well as the addition of reinforcement learning to the topology search. </br> [View full release notes](https://convstruct.org)
<br>
—
<img align=”right” width=”250” height=”250” src=”https://i.ibb.co/x8WJmPS/convstruct-icon.png”>
<br>
Convstruct is an open source Python framework, containing four functions to create your own topology search for:
• Computer vision models <br> • NLP/NLU models <br> • Reinforcement learning models
<br> <br>
—
<br>
Convstruct, cs, seamlessly works alongside Tensorflow, tf, when your creating your AI models. Code your loss function and optimizer and let Convstruct do the rest.
<img align=”center” src=”https://i.ibb.co/QjHf2vv/main-2.png” alt=”convstruct-functions-banner” border=”0”>
<br>
—
<br>
###Getting started with the python package?
• Install using pip: pip install convstruct <br> • Import Convstruct inside your python file.
###Getting started with the codebase?
• Clone this repo to your local machine using https://github.com/convstruct/convstruct <br> • Run pip install requirements.text in the cloned repo to make sure all Convstruct dependencies are installed on your local machine. <br> • Import Convstruct inside your python file.
> Convstruct 1.1 supports only Python 3.6 and Tensorflow 1.14
<br>
—
###Convstruct roadmap The roadmap provides a high-level visual of what has been done and what is planned next for Convstruct.
<img align=”center” src=”https://i.ibb.co/RCcntFy/main-3.png” alt=”convstruct-roadmap-banner” border=”0”>
<br> <br>
—
<br>
For inquiries into Convstruct contact: [hello@convstruct.org](hello@convstruct.org)
<br> <br>
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for convstruct-1.1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 563cb2aa0e6b683784d384c4f9b76c804202a93ea3bdedbc977ed05336731b9a |
|
MD5 | 1f10322a835b4f223a41cb47b5aebb07 |
|
BLAKE2b-256 | 6d5959b3807eaba20dd975713afa5af1d0469a37349dd8344cad381437b5ded2 |