The Machine Learning Hub
========================
A *Command Line* framework and platform for presenting and utilising
Machine Learning, Artificial Intelligence, and Data Science
capabilities.
Introduction
------------
The [machine learning hub](https://mlhub.ai) is a free and open source
project hosted on [github](https://github.com/mlhubber/mlhub) aimed at
easily sharing [machine learning, artificial intelligence, and data
science models and technologies](https://github.com/mlhubber/mlmodels). The
source code for such models will be found on
[github](https://github.com), [gitlab](https://gitlab.com), or
[bitbucket](https://bitbucket.org). The models are installed and
managed using the *ml* command from the *mlhub* package designed to
install the model and run a demonstration within 5 minutes, as well as
providing a suite of useful command line tools based on the pre-built
models. Each model has been tested on Ubuntu (GNU/Linux).
A number of demonstration models have been packaged and are listed in
the [repository index](https://mlhub.ai/Packages.html) on
[mlhub.ai](https://mlhub.ai/) where the models themselves can be
browsed.
In this blog post we first review how to get started with MLHub and
then illustrate some of its functionality. See the growing
documentation at the [GNU/Linux Survival
Guide](https://togaware.com/linux/survivor/AI_Machine.html).
Quick Start
-----------
The command line interface can be installed using
[PyPi](https://pypi.org/project/mlhub/):
$ pip3 install mlhub
Once installed you will be able to run any curated or other model. The quick
start version is below. Note that you type the command `ml ...` and
that everything from the `#` to the end of the line is ignored (it's
a comment):
$ ml install <model> # Install the named model.
$ ml demo <model> # Run the demonstration of the model.
$ ml gui <model> # Graphical display to utilise the mode.
The following commands are available and below is a brief description of
each command:
$ ml # Show a usage message.
$ ml available # List of pre-buld models on the MLHub.
$ ml installed # List of pre-built models installed locally
$ ml install <model> # Install the model named 'rain'.
$ ml configure <model> # Install required dependencies.
$ ml readme <model> # View background information about the model.
$ ml commands <model> # List of commands supported by the model.
$ ml demo <model> # Run the demonstration of the model
$ ml gui <model> # Graphical display of pre-built model.
$ ml score <model> # Run model on your own data.
Different model packages will have different dependencies and these will
be installed by the *configure* command.
Quick Start: Azure DSVM
-----------------------
A particularly attractive and simple way to get started with exploring
the mlhub functionality is to fire up a [Ubuntu Data Science Virtual
Machine](https://aka.ms/dsvm) (DSVM) on Azure for as little as USD10 per
month for quite a small server or USD90 for a reasonable one. You can
get free credit (USD200) from Microsoft to [trial the
DSVM](https://aka.ms/free).
Using this virtual machine will save a lot of time compared with setting
up your own machine with the required dependencies, which of course you
can do if you wish as all the dependencies are open source.
To set up the virtual machine, with an Azure subscription log in to the
[portal](https://portal.azure.com/) and add a new Data Science Virtual
Machine for Linux (Ubuntu). You need to provide a name (for the virtual
machine), a user name and a password, and then create a new resource
group and give it a name, and finally choose a location. Go with all the
defaults for everything else, except choose a size to suit the budget
(B1s is cheap though a D2s is a better interactive experience). Note
that you are only charged whilst the machine is fired up so USD90 per
month is no where near what you will spend if you only fire up the
server when you need.
Once the DSVM is set up go to its Overview page and click on DNS name
Configure and provide a name by which to refer to the server publicly
(e.g., myml.westus2.cloudapp.azure.com).
We now have a server ready to showcase the pre-built Machine Learning
models. There are several options to connect to the server but a
recommended one is to use [X2Go](https://x2go.org/) which supports
Linux, Windows, and Mac. Install it and point it to your server (e.g.,
myml.westus2.cloudapp.azure.com) in the setup.
Connect to the DSVM. Close the Firefox window that pops up. Click on the
terminal icon down the bottom, and you are ready to go:
$ pip install mlhub
$ ml
$ ml available
etc.
Pre-Built Model Archives
------------------------
A model is a zip file archived as .mlm files and hosted in a repository.
The public repository is [mlhub.ai](https://mlhub.ai/). The *ml* command
can install the pre-built model locally, ready to run a demo, to print
and display the model, and to score new data using the model. Some
models provide ability to retrain the model with user provided data.
Contributing Models to ML Hub
-----------------------------
Anyone is welcome to contribute a pre-built model package to ML Hub.
Please submit a pull request through
[github](https://github.com/mlhubber).
Installing Pip3
---------------
On Ubuntu this is as simple as:
$ sudo apt install python3-pip
Alternative pip Install
-----------------------
Depending on your setup of pip, you may need to use:
$ pip3 install mlhub
The executable may be placed into `~/.local/bin` which will need to be
on your path. Edit your shell startup which is either `.profile` or
`.bashrc`, etc:
PATH="$HOME/.local/bin:$PATH"
Alternative Install
-------------------
A tar.gz containing the mlhub package and the command line interface is
available as
[mlhub_3.7.0.tar.gz](https://mlhub.ai/dist/mlhub_3.7.0.tar.gz) within
the [distribution folder](https://mlhub.ai/dist/) of the MLHub.
To install from the tar.gz file:
$ wget https://mlhub.ai/dist/mlhub_3.7.0.tar.gz
$ pip install mlhub_3.7.0.tar.gz
$ ml
Or extract the above downloaded .tar.gz and install:
$ wget https://mlhub.ai/dist/mlhub_3.7.0.tar.gz
$ tar xvf mlhub_3.7.0.tar.gz
$ cd mlhub
$ python3 setup.py install --user
Under Development
-----------------
An interactive MLHub session that is initiated through the demo
command is quite similar to a Jupyter Notebook presentation running on
top of a Jupyter interpreter. Notebooks can be automatically
transformed into a MLHub package so that the notebook becomes the
source for the interactive demo.py or demo.R script required by
MLHub. In this way users have the choice to either run the Notebook
interactively within Jupyter or from the command line as an
interactive script.
Contributions
-------------
The open source mlhub command line tool (ml) and sample models are being
hosted on [github](https://github.com/mlhubber) and contributions to
both the command line tool and contributions of new open source
pre-built machine learning models are most welcome. Feel free to submit
pull requests.
Metrics
-------
MLHub PyPI download statistics: https://pepy.tech/project/mlhub
[](https://pepy.tech/project/mlhub)
[](https://pepy.tech/project/mlhub)
[](https://pepy.tech/project/mlhub)
MLHub Dev PyPI download statistics: https://pepy.tech/project/mlhubdev
[](https://pepy.tech/project/mlhubdev)
[](https://pepy.tech/project/mlhubdev)
[](https://pepy.tech/project/mlhubdev)
Raw data
{
"_id": null,
"home_page": "https://mlhub.ai",
"name": "mlhub",
"maintainer": null,
"docs_url": null,
"requires_python": null,
"maintainer_email": null,
"keywords": "machine learning models repository",
"author": "Graham Williams",
"author_email": "mlhub@togaware.com",
"download_url": "https://files.pythonhosted.org/packages/f2/de/ed532446553e1965764b3fcb16b8ddf8678d944641d7bcd1f826389716dd/mlhub-3.11.3.tar.gz",
"platform": null,
"description": "The Machine Learning Hub\n========================\n\nA *Command Line* framework and platform for presenting and utilising\nMachine Learning, Artificial Intelligence, and Data Science\ncapabilities.\n\nIntroduction\n------------\n\nThe [machine learning hub](https://mlhub.ai) is a free and open source\nproject hosted on [github](https://github.com/mlhubber/mlhub) aimed at\neasily sharing [machine learning, artificial intelligence, and data\nscience models and technologies](https://github.com/mlhubber/mlmodels). The\nsource code for such models will be found on\n[github](https://github.com), [gitlab](https://gitlab.com), or\n[bitbucket](https://bitbucket.org). The models are installed and\nmanaged using the *ml* command from the *mlhub* package designed to\ninstall the model and run a demonstration within 5 minutes, as well as\nproviding a suite of useful command line tools based on the pre-built\nmodels. Each model has been tested on Ubuntu (GNU/Linux).\n\nA number of demonstration models have been packaged and are listed in\nthe [repository index](https://mlhub.ai/Packages.html) on\n[mlhub.ai](https://mlhub.ai/) where the models themselves can be\nbrowsed.\n\nIn this blog post we first review how to get started with MLHub and\nthen illustrate some of its functionality. See the growing\ndocumentation at the [GNU/Linux Survival\nGuide](https://togaware.com/linux/survivor/AI_Machine.html).\n\nQuick Start\n-----------\n\nThe command line interface can be installed using\n[PyPi](https://pypi.org/project/mlhub/):\n\n $ pip3 install mlhub\n\nOnce installed you will be able to run any curated or other model. The quick\nstart version is below. Note that you type the command `ml ...` and\nthat everything from the `#` to the end of the line is ignored (it's\na comment):\n\n $ ml install <model> # Install the named model.\n $ ml demo <model> # Run the demonstration of the model.\n $ ml gui <model> # Graphical display to utilise the mode.\n\nThe following commands are available and below is a brief description of\neach command:\n\n $ ml # Show a usage message.\n $ ml available # List of pre-buld models on the MLHub.\n $ ml installed # List of pre-built models installed locally\n $ ml install <model> # Install the model named 'rain'.\n $ ml configure <model> # Install required dependencies.\n $ ml readme <model> # View background information about the model.\n $ ml commands <model> # List of commands supported by the model.\n $ ml demo <model> # Run the demonstration of the model\n $ ml gui <model> # Graphical display of pre-built model.\n $ ml score <model> # Run model on your own data.\n\nDifferent model packages will have different dependencies and these will\nbe installed by the *configure* command.\n\nQuick Start: Azure DSVM\n-----------------------\n\nA particularly attractive and simple way to get started with exploring\nthe mlhub functionality is to fire up a [Ubuntu Data Science Virtual\nMachine](https://aka.ms/dsvm) (DSVM) on Azure for as little as USD10 per\nmonth for quite a small server or USD90 for a reasonable one. You can\nget free credit (USD200) from Microsoft to [trial the\nDSVM](https://aka.ms/free).\n\nUsing this virtual machine will save a lot of time compared with setting\nup your own machine with the required dependencies, which of course you\ncan do if you wish as all the dependencies are open source.\n\nTo set up the virtual machine, with an Azure subscription log in to the\n[portal](https://portal.azure.com/) and add a new Data Science Virtual\nMachine for Linux (Ubuntu). You need to provide a name (for the virtual\nmachine), a user name and a password, and then create a new resource\ngroup and give it a name, and finally choose a location. Go with all the\ndefaults for everything else, except choose a size to suit the budget\n(B1s is cheap though a D2s is a better interactive experience). Note\nthat you are only charged whilst the machine is fired up so USD90 per\nmonth is no where near what you will spend if you only fire up the\nserver when you need.\n\nOnce the DSVM is set up go to its Overview page and click on DNS name\nConfigure and provide a name by which to refer to the server publicly\n(e.g., myml.westus2.cloudapp.azure.com).\n\nWe now have a server ready to showcase the pre-built Machine Learning\nmodels. There are several options to connect to the server but a\nrecommended one is to use [X2Go](https://x2go.org/) which supports\nLinux, Windows, and Mac. Install it and point it to your server (e.g.,\nmyml.westus2.cloudapp.azure.com) in the setup.\n\nConnect to the DSVM. Close the Firefox window that pops up. Click on the\nterminal icon down the bottom, and you are ready to go:\n\n $ pip install mlhub\n $ ml\n $ ml available\n\netc.\n\nPre-Built Model Archives\n------------------------\n\nA model is a zip file archived as .mlm files and hosted in a repository.\nThe public repository is [mlhub.ai](https://mlhub.ai/). The *ml* command\ncan install the pre-built model locally, ready to run a demo, to print\nand display the model, and to score new data using the model. Some\nmodels provide ability to retrain the model with user provided data.\n\nContributing Models to ML Hub\n-----------------------------\n\nAnyone is welcome to contribute a pre-built model package to ML Hub.\nPlease submit a pull request through\n[github](https://github.com/mlhubber).\n\nInstalling Pip3\n---------------\n\nOn Ubuntu this is as simple as:\n\n $ sudo apt install python3-pip\n\nAlternative pip Install\n-----------------------\n\nDepending on your setup of pip, you may need to use:\n\n $ pip3 install mlhub\n\nThe executable may be placed into `~/.local/bin` which will need to be\non your path. Edit your shell startup which is either `.profile` or\n`.bashrc`, etc:\n\n PATH=\"$HOME/.local/bin:$PATH\"\n\nAlternative Install\n-------------------\n\nA tar.gz containing the mlhub package and the command line interface is\navailable as\n[mlhub_3.7.0.tar.gz](https://mlhub.ai/dist/mlhub_3.7.0.tar.gz) within\nthe [distribution folder](https://mlhub.ai/dist/) of the MLHub.\n\nTo install from the tar.gz file:\n\n $ wget https://mlhub.ai/dist/mlhub_3.7.0.tar.gz\n $ pip install mlhub_3.7.0.tar.gz\n $ ml\n\nOr extract the above downloaded .tar.gz and install:\n\n $ wget https://mlhub.ai/dist/mlhub_3.7.0.tar.gz\n $ tar xvf mlhub_3.7.0.tar.gz\n $ cd mlhub\n $ python3 setup.py install --user\n\nUnder Development\n-----------------\n\nAn interactive MLHub session that is initiated through the demo\ncommand is quite similar to a Jupyter Notebook presentation running on\ntop of a Jupyter interpreter. Notebooks can be automatically\ntransformed into a MLHub package so that the notebook becomes the\nsource for the interactive demo.py or demo.R script required by\nMLHub. In this way users have the choice to either run the Notebook\ninteractively within Jupyter or from the command line as an\ninteractive script.\n\nContributions\n-------------\n\nThe open source mlhub command line tool (ml) and sample models are being\nhosted on [github](https://github.com/mlhubber) and contributions to\nboth the command line tool and contributions of new open source\npre-built machine learning models are most welcome. Feel free to submit\npull requests.\n\nMetrics\n-------\n\nMLHub PyPI download statistics: https://pepy.tech/project/mlhub\n\n[](https://pepy.tech/project/mlhub)\n[](https://pepy.tech/project/mlhub)\n[](https://pepy.tech/project/mlhub)\n\nMLHub Dev PyPI download statistics: https://pepy.tech/project/mlhubdev\n\n[](https://pepy.tech/project/mlhubdev)\n[](https://pepy.tech/project/mlhubdev)\n[](https://pepy.tech/project/mlhubdev)\n\n",
"bugtrack_url": null,
"license": null,
"summary": "Machine learning model repository manager",
"version": "3.11.3",
"project_urls": {
"Homepage": "https://mlhub.ai"
},
"split_keywords": [
"machine",
"learning",
"models",
"repository"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "f2deed532446553e1965764b3fcb16b8ddf8678d944641d7bcd1f826389716dd",
"md5": "d90d53d894c295a0896e2cec6a1298ec",
"sha256": "62da1d405e8c553911339894bec3196ec19e89a69feb2ff5c2c7eda8f021f915"
},
"downloads": -1,
"filename": "mlhub-3.11.3.tar.gz",
"has_sig": false,
"md5_digest": "d90d53d894c295a0896e2cec6a1298ec",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 57695,
"upload_time": "2024-12-05T21:53:17",
"upload_time_iso_8601": "2024-12-05T21:53:17.265151Z",
"url": "https://files.pythonhosted.org/packages/f2/de/ed532446553e1965764b3fcb16b8ddf8678d944641d7bcd1f826389716dd/mlhub-3.11.3.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-05 21:53:17",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "mlhub"
}
JSON