It's been two months since this profile received a new review

MLlib Reviews & Product Details

MLlib is Spark's machine learning (ML) library that make practical machine learning scalable and easy it provides ML Algorithms: common learning algorithms such as classification, regression, clustering, and collaborative filtering, feature extraction, transformation, dimensionality reduction, and selection, tools for constructing, evaluating, and tuning ML Pipelines, saving and load algorithms, models, and Pipelines and linear algebra, statistics, data handling, etc.

Seller

The Apache Software Foundation

Product Description

Value at a Glance

Averages based on real user reviews.

Time to Implement

1 month

View More Pricing Information

Top-Rated Alternatives

View All Alternatives

MLlib Reviews (14)

G2 reviews are authentic and verified.

Here's how.

Chetan S.

Data Analyst

Small-Business (50 or fewer emp.)

10/10/2020

"Apache Spark - MLib review"

4/5

What do you like best about MLlib?

It is useful in implementing machine learning algorithms like classification, regression and clustering. It works well while using statistical modelling techniques Review collected by and hosted on G2.com.

What do you dislike about MLlib?

It has an expensive memory with the necessity of manual optimization which might degrade user experience. It gives latency but can be used amongst R and python communities Review collected by and hosted on G2.com.

Recommendations to others considering MLlib:

This can be preferred if the request is to extract and access the data quickly. Also certain algorithms work well with the tool based upon the distinct requirements. Budget is also a factor to be looked upon Review collected by and hosted on G2.com.

What problems is MLlib solving and how is that benefiting you?

ETL and data extraction. Fast data accessing can be performed using the tools Review collected by and hosted on G2.com.

Mohini S.

Small-Business (50 or fewer emp.)

10/10/2020

"MLlib review"

4/5

What do you like best about MLlib?

implementation of ML algorithms like regression, classification and modelling techniques can be done using the tool Review collected by and hosted on G2.com.

What do you dislike about MLlib?

MLlib is not production ready, moreover Spark does not come out as a useful engine owing to its latency Review collected by and hosted on G2.com.

Akshay K.

Data Analyst

Mid-Market (51-1000 emp.)

10/9/2020

"Great Software!"

5/5

What do you like best about MLlib?

The interface and the workstation is to top notch. Easy to navigate and experiment with. Review collected by and hosted on G2.com.

What do you dislike about MLlib?

Nothing at all. All are perfect and efficient enough. Review collected by and hosted on G2.com.

Kunal B.

Senior Engineer - Data Engineering

Mid-Market (51-1000 emp.)

8/28/2020

"Effectiveness of Mlib"

5/5

What do you like best about MLlib?

Distributed computing helps in speed and efficiency Review collected by and hosted on G2.com.

What do you dislike about MLlib?

Nothing is bad, everything about Spark is great Review collected by and hosted on G2.com.

Verified User in Financial Services

Mid-Market (51-1000 emp.)

12/10/2019

"Best scalable machine learning framework."

4/5

What do you like best about MLlib?

The scalability power of the framework which handles large data efficiently and performs machine learning algorithms at faster rate. Review collected by and hosted on G2.com.

What do you dislike about MLlib?

The syntax and code changes for python R depends on the tools we are using.It is not standard which is tough for new users to adapt.The packages are very different compared tools to tool. Review collected by and hosted on G2.com.

Recommendations to others considering MLlib:

If your problem is the large data to solve organization problems using machine learning then MIlb is the right one to use. Review collected by and hosted on G2.com.

What problems is MLlib solving and how is that benefiting you?

We are solving the large data problems in our organization so that it would be salable and works faster for us. Review collected by and hosted on G2.com.

Dhawal G.

Undergraduate Reseacher , Mechatronics Instrumentation and Control Lab

Research

Small-Business (50 or fewer emp.)

2/19/2019

"ML Lib a Machine Learning library on Spark"

4/5

What do you like best about MLlib?

MLLib was used as part of course in my college for Big Data. So we got to study why actually mllib came about and what all inadequacies were there in the Map-Reduce Framework of Hadoop and how apache Spark has solved them. The best part is the ease of use of Mllib and also the excellent documentation support from both the official website as well as the sources outside like youtube videos. The big community makes it easy to learn and use mllib. I used mllib for decision trees and I being a student was successfully able to implement the same with ease. Plus the python implementation is very easy to implement. Review collected by and hosted on G2.com.

What do you dislike about MLlib?

We were given a preinstalled system for our labs and a cluster, but when I tried to do the same for my machine, I found it rather tricky to install. Also, support for deep learning is not there, which is a very fast growing field of machine learning. Review collected by and hosted on G2.com.

Recommendations to others considering MLlib:

Good and easy to use library for multi cluster computing but only for conventional machine learning problems. Currently not adept with the deep learning support which may be nice in the future. Review collected by and hosted on G2.com.

What problems is MLlib solving and how is that benefiting you?

I did a course on Big Data where I used Hadoop and apache spark to learn the various techniques used to deal with big data. Here I used MLlib to do a course project on classification, where I built a model a decision tree model from the data that I acquired by scraping humongous amount of sites. Review collected by and hosted on G2.com.

Verified User in Telecommunications

Enterprise (> 1000 emp.)

12/2/2019

"A good library with futuristic short comings "

3.5/5

What do you like best about MLlib?

MLib so far is the best community supported widely used machine learning library for apache spark Review collected by and hosted on G2.com.

What do you dislike about MLlib?

MLib is inconsistent with deep learning models, this causes issues while moving models to production Review collected by and hosted on G2.com.

Verified User in Computer Software

Mid-Market (51-1000 emp.)

12/2/2019

"Useful tool for in-memory ML pipelines"

3.5/5

What do you like best about MLlib?

Speed and ease of use. Strong community support and lots of resources. Review collected by and hosted on G2.com.

What do you dislike about MLlib?

Prototyping can be time consuming. Also, limited utility in case of extremely large datasets. Review collected by and hosted on G2.com.

Verified User in Computer Software

Mid-Market (51-1000 emp.)

3/21/2019

"MLlib is a convenient parallelized ML library "

4.5/5

What do you like best about MLlib?

I love how it includes most of the popular ML libraries for easy use with Apache Spark and parallelized computing. The power is only limited by the number of cores you've got. Review collected by and hosted on G2.com.

What do you dislike about MLlib?

I feel like some other ML frameworks have better models, or added features/functionality used in developing models. MLlib is also an open source part of Spark, so development of the framework depends largely on what Open Source folks contribute to it. Review collected by and hosted on G2.com.

What problems is MLlib solving and how is that benefiting you?

I'm doing ML problems with Apache Spark dataframes. The benefits are we can massively parallelize our training and modeling. I've worked with customers who used MLlib to build out random forest decision trees with massive tree depth and massive tree count. This would be impossible without MLlib. Review collected by and hosted on G2.com.

Saeid A.

Data Scientist and Researcher

Outsourcing/Offshoring

Enterprise (> 1000 emp.)

4/20/2018

"Distributed ML in Spark with limited flexibility, especially for advanced algorithms"

2.5/5

What do you like best about MLlib?

It is distributed and allow distributed execution of model training ans well as model scoring. It helps to leverage benefit of Spark without using Scala. It delivers Spark ML with Python!

High performance since it is a RDD-based data modeling package.

Fairly nice documentation. Review collected by and hosted on G2.com.

What do you dislike about MLlib?

It is rigid with some of the algorithms, specially with advanced one like neural network. For instance, you are unable to change activation functions of a neural network. You can either use Sigmoid for all the layers, or tanh which is not really making sense!

Evaluation metrics are not as rich as packages like Scikit-Learn.

Not all its functionalities implemented in Python. Many are Scala-based yet. Review collected by and hosted on G2.com.

Recommendations to others considering MLlib:

If you bother about advanced algorithm in specific neural network, do not use MLlib as it does give you least flexibility in customizing the network.

Perhaps it is great for regression and decision tree in distributed environment. Review collected by and hosted on G2.com.

What problems is MLlib solving and how is that benefiting you?

MLlib has both classification and regression algorithms under supervised learning and also k-means under unsupervised learning.

The beauty of the package lies in its distributed execution. Review collected by and hosted on G2.com.