FloydHub Reviews

It has allowed us to start looking/building software for future use

It has allowed some of our colleagues to get distracted

none at this time

downloading images directly much faster then before

It's easy to set up, you can train your models in an easy way and that's what I appreciate about this software. The customer service is responsive and I like that they answer fastly.

I always have issues uploading big amounts of data but that is something that could be improved.

it will help you do your job more easy!

It makes our job easier and we are able to train out models and resolve our problems.

With FloydHub it's simple to transition from building deep networks on my computer to training them on GPUs. Maybe the best part is that nearly every environment I need comes pre-packaged so no additional setup is necessary, which isn't always the case with other cloud GPU solutions. When I require other libraries, I don't have to jump through hoops to install them. Finally, they provide cheap storage so it's easy to create large datasets and attach them to an instance, again with almost no additional setup.

Nothing. Great product and customer support. In fact, once when I accidentally left a Jupyter Notebook running all weekend, they partially refunded me the GPU hours. Then, so it wouldn't happen again, they went the extra mile, assuring me that in the future any Jupyter Notebook left idle for more than an hour would be shut down automatically. Pretty amazing, considering it was my fault.

Since deep learning often requires lots of finetuning, the platform's ease of use allows for quick experimentation that gives me the means to design better models.

The API is straightforward, well documented, and makes it easy to prototype and test models with a minimum of outside setup. They also offer a fairly large range of instance images which cover every deep learning framework that I have used and some that I have not. The team also has been extremely responsive to concerns, and fixed several major pain points when they were brought up. This makes me confident in both the team and the future of the product .

I don't think it's an efficient alternative yet for large projects- it's great for experimentation, testing, and small-to-medium size projects. If you are doing research, experimentation, or something along those lines it is perfect.

If you are looking for the flexibility, ease of setup and use, then Floydhub is a great alternative to spinning up an instance. The runway for me from learning about the software to running my first job was less than an hour.

It makes it easy to test, prototype, and explore modelling options.

I really appreciate the simplicity of Floydhub to train my models, you only need one command line to trigger a Job. This is a game changer compared to using AWS GPU instances directly:

- you don't have to worry about the storage of your datasets

- you don't have any hidden fees and you pay computation per minute (instead of hourly)

- each train you did is versioned on Floydhub

- Training jobs starts really quickly

- Jupyter/Tensorboard support

The only issue I had with Floydhub is some instability when downloading Job outputs, but it has been fixed. Floydhub is under development and may have some bugs or missing features, but it's a very promising service!

It greatly simplifies deep learning training pipeline, which enables me to focus on the algorithms, you won't need DevOps with Floydhub (Heroku for Deep learning).

Responsive to messages when I need help. The platform has most of the things I need, yet is extendable where it needs to be - for instance it enables installing missing python packages when running the code.

There are some minor UI tweaks that I would suggest, but I already did and they are working on it.

If you need to run deep learning models on GPU without much configuration, this is the place for you. There is also trial version when you register so you can try it out before paying anything.

Running deep learning models on GPUs, which can be essential if your business is using machine learning.

It's easy to gear up and get started. I spent 2 weeks earlier this year, trying to get set up on AWS. I think that's gotten better, but FloydHub was exponentially easier.

Nothing so far. They've been very responsive. With every experiment, I'm getting deeper and deeper into how flexible they are for my Deep Learning work.

Unless you're an AWS wizard, you should definitely give FloydHub a try. For its ease of use - everything you need is available easy to spin up. The cost is also much better. AWS seemed to charge me, even when I was doing nothing. Also, FloydHub will automatically shut down your project if you forget to close it. AWS does not and you can incur HUGE charges if you're not careful. Also, Floydhub lets you use their service by the minute (instead of how AWS charges you for an hour, even if you use a few minutes.) The per hour charge is about half of AWS.

I'm in an experimental phase. I am trying to used ML/DL/NLP to make Advertising Creative better. I'm an advertising copywriter by trade, which is why the ease of FloydHub is so appealing to me.

Best thing is that it lets you learn and is an overall great machine learning hub. For beginners, you can create unlimited public projects for no price at all. Their Cloud GPUs are great, and so is their version control and training modules. If you actually go for the paid version you also get to create private data sets, at a very reasonable price. The dashboard is quite effective, as well.

I do not like the amount of space provided, as I feel that it is too low. Considering how a simple google drive can offer way more space than that, I would like for them to increase the size they provide for being able to store bigger projects.

I do recommend storing your projects and codes on this, because the deployment is very easy through this.

I often use it to manage my projects and deploy trained models. It is good for researching your materials and controlling all your data.

Like the continues multi simulated services and ensuring services is fully running. As well the version control that is necessary for cloud base deployment

We encountered in one of our service instance that still running even though its reach the limit that we set and they assisted us for the proper configuration of setting up the limit of the instance that will be beneficial for us.

If you need low cost but effecient this the product you should get and they have great customer assistance that will help you out if there is any issue/question you have.

It gave us more options and service that suitable in our needs within the budget.

It is very helpful for team collaboration. The best feature is that you can run multiple projects at the same time and work on them simultaneously. it is easier to compare prices and the general projects and generate necessary data from that. Their cloud GPUs are pretty good and the price of the software is pretty reasonable as well.

Not much of cloud storage available. Usually i can only store a few projects in it and often have to store them somewhere else to reload new projects. Does not work on Linux.

I would recommend using it for storing chunks of individual projects and not for backing up every single thing.

I use it to store our projects and also manage a tiny chunk of the database through it. It is easy to use and has a lot of nice features available to use.