Summary of Automl Solutions – List and Comparison – Dan Rose AI

presentation

I have been looking for a list of vehicle solutions and a way to compare them, but I was unable to find it. So I thought that I too could compile that list for others to use it. If you are not familiar with the car, read this post for a quick presentation and the good and the bad.

I have not been able to try everyone and make a proper compilation, so this is just a feature -based comparison. I tried to choose the features that felt the most important to me, but it may not be the most important to you. If you think that some traits are missing or if you know an automobile solution that should be on the list just let me know.

Before we go to the list, I would simply quickly go through features and how I interpret them.

FeatureS

SETTLING

Some solutions can be automatically placed directly on the Cloud with a one -click setting. Some simply export to Tenzorflow and some even have specific export to the skirt devices.

The types

This can be text, images, videos, tables. I think some of the open sources can lie down to do nothing if put to work, so it may not be the full truth.

Explicable

Explanation in it is a hot topic and a very important feature for some projects. Some solutions do not give you knowledge and some give you a lot and can even be a strategic differentiation for the provider. I have simply divided this feature into few, some and very explainable.

Monitor

Monitoring patterns after setting to avoid moving patterns can be a very useful feature. I shared this in yes and no.

Accessible

Some of the providers are very easy to use and some of them require coding and at least a fundamental understanding of data science. So I took this feature so that you can choose the tool that corresponds to the skills you have access to.

Tag

Some have an internal labeling tool so that you can directly label the data before training the model. This can be very useful in some cases.

General / specialized

Most vehicle solutions are generalized for all industries, but some are specialized for specific industries. I suspect this will become more popular, so I got this feature.

Open -source

Self-explanatory. Is it an open source or not.

Includes the transfer of transfer

Transfer learning is one of the great advantages of the Automl. You go to the piggyback on large patterns in order to get great results with very little data.

List of vehicle solutions

Google Auto

Google Automl is the one I am most popular with. I saw it easy enough to use even without coding. The biggest iswash I have had is that API requires a configuration bunch and is not just a simple certificate or OAUTH -based authentication.

Setting: In cloud, export, edge

Types: Text, images, videos, table

Explainable: Small

Monitor: not

Accessible: many

Tool of labeling: Used to have but is closed

General / specialized: Generalized

Open Source: not

Includes transfer lesson: yes

Link: link https://cloud.google.com/automl

Azure

Microsoft’s Cloud Automl seems to be larger than Google but with only table data models.

Setting: In cloud, some local

Types: Just the table

Explainable: some

Monitor: not

Accessible: many

Tool of labeling: not

General / specialized: Generalized

Open Source: not

Includes transfer lesson: yes

Link: link https://azure.microsoft.com/en-us/rvice/machine-learning/automatedml/

Lobe.ai

This solution is still in beta, but it works very well in my experience. I will write a summary as soon as possible to go public. The lobby is so easy to use that you can allow a 10-year-old to use it to train deep learning models. I would really recommend this for educational purposes.

Setting: Local and export to Tenororflow

Types: imaging

Explainable: Small

Monitor: –

Accessible: Many – a third class can use this

Tool of labeling: yes

General / specialized: Generalized

Open Source: not

Includes transfer lesson: yes

Link: link https://lobe.ai/

Cortical

Cortical seems to be one of the solutions of vehicles that distinguish themselves by being as explained as possible. This can be a great advantage when not only trying to get good results, but also understand the business problem better. For this I am a little worshiper.

Setting: Newly

Types: tabular

Explainable: many

Monitor: not

Accessible: many

Tool of labeling: not

General / specialized: Generalized

Open Source: not

Includes transfer lesson: Not sure

Link: link https://kortical.com/

Computer robot

A big player who can even be the first clean vehicle that goes IPO.

Setting: Newly

Types: Text, images and tables

Explainable: many

Monitor: yes

Accessible: many

Tool of labeling: not

General / specialized: Generalized

Open Source: not

Includes transfer lesson: yes

Link: link https:

AWS Sagemaker Autopilot

Amazons Automl. Requires more technical skills than other large Cloud suppliers and is quite limited and supports only two algorithms: XGBOost and logistical regression.

Setting: In the cloud and export

Types: tabular

Explainable: some

Monitor: yes

Accessible: Coding

Tool of labeling: yes

General / specialized: Generalized

Open Source: not

Includes transfer lesson: yes

Link: link https://aws.amazon.com/sagemaker/autopilot/

waist

Setting: Export and cloud

Types: tabular

Explainable: yes

Monitor: –

Accessible: many

Tool of labeling: not

General / specialized: Generalized

Open Source: Mljar has both open sources (https://github.com/mljar/mljar-supervised) and closed source solutions.

Includes transfer lesson: yes

Link: link https://mljar.com/

Autogluon

Setting: export

Types: Text, images, tables

Explainable: –

Monitor: –

Accessible: Coding

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson: yes

Link: link https://autogluon.mxnet.io/

Jadbio

Setting: Cloud and export

Types: tabular

Explainable: some

Monitor: not

Accessible: many

Tool of labeling: not

General / specialized: longevity

Open Source: not

Includes transfer lesson: –

Link: link https://www.jadbio.com/

car

This solution supports Bayesian models which is quite delightful.

SETTLING : Export

Types: –

Explainable: –

Monitor: –

Accessible: Code

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson:not

Link: link https://www.cs.ubc.ca/labs/beta/projecs/autoweka/

H2o he saw the driver

Also supports Bayesian models

Setting: export

Types: –

Explicable: – –

Monitor: –

Accessible: Half

Tool of labeling: not

General / specialized: Generalized

Open Source: Both options

Includes transfer lesson: –

Link: link https://www.h2o.ai/

Self -proclaimer

Autokara is one of the most popular open source solutions and is definitely worth trying.

Setting: export

Types: Text, images, tables

Explainable: Potential

Monitor: –

Accessible: Code

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson: –

Link: link https://autokeras.com/

Tip

Setting: export

Types: Images and tables

Explainable: Potential

Monitor: –

Accessible: Code

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson: –

Link: link http://epistasislab.github.io/tpot/

Picaret

Setting: export

Types: Text, table

Explainable: Potential

Monitor: –

Accessible: Code

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson: –

Link: link https://github.com/pycaret/pycaret

collection

Setting: export

Types: tabular

Explainable: Potential

Monitor: –

Accessible: Code

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson: –

Link: link https://automl.github.io/auto-klearn/master/

Transmogrifai

Made from Salesforce.

Setting: export

Types: Text and table

Explainable: Potential

Monitor: –

Accessible: Code

Tool of labeling: not

General / specialized: Generalized

Open Source: yes

Includes transfer lesson: –

Link: link https://transmogrif.ai/