Google Cloud AutoML

What do you like best about Google Cloud?

That has a lot of services and I especially like the logging.

What do you dislike about Google Cloud?

The documentation can be chaotic and daunting. I sometimes loose myself there.

What problems is Google Cloud solving and how is that benefiting you?

The deployment of my models, the execution of them and also the data hosting comes from google. It helps to have all the tools I need in one place and they can interact with each other without issues and more configuration from my side.

What do you like best about Google Cloud AutoML?

Google AutoML kit is one of the best platforms for developers of Machine Learning, as the main benefit is that it is from Google and ML is highly affected by Google. In that manner, Google's AutoML kit provides very sophisticated ML techniques and algorithms support. In addition to that, the performance and accuracy of AutoML algorithms are way better than other ML platforms. Create and train the model of any specific problems is very easy and fast. Google AutoML Algorithms and techniques can easily handle large and complex data processing and as is it on the cloud, I get extremely good GPU support that would rather cost me thousands of bucks. It really makes the difference in how I develop ML applications.

What do you dislike about Google Cloud AutoML?

One thing I believe can be improved that Google can get know regarding my private data sets which sometimes can be a privacy issue for some users. So, Google may introduce some private data loading techniques on AutoMl platform. Other than that there are not any major issues in it.

Recommendations to others considering Google Cloud AutoML:

Google Could AutoML is a very cost-effective solution in terms of hardware resources. Also, with that, you will get security as well as reliable performance for any kind of complex data you want to process. Definitely anyone who is working in ML should go with it once.

What problems is Google Cloud AutoML solving and how is that benefiting you?

I use Google Could AutoML for various purposes such as data preprocessing, dataset manipulation and mainly for model training. We have monthly subsciption based plan for Google AutoML platform at our company and as a machine learning developer, I am utilizing it for developing a different pilot application for demo purpose as well as main projects of ML in it.

What do you like best about Google Cloud AutoML?

This helped our company improve decision-making more quickly and easily. Faced with a set of data of interest, AutoML extracts the dataset of interest from users and determines a high-performance network architecture. It is the quick and easy way to start working quickly and obtain sophisticated models later. Ease of use and the ability to hire interns to tag photos so the senior team can focus on technology.

What do you dislike about Google Cloud AutoML?

There is still a lot of art associated with training machine learning models. Difficult for us to debug bad image matches. Every now and then the beta has a flaw that I can't remove from my model.

What problems is Google Cloud AutoML solving and how is that benefiting you?

I was trying to build an image recognition app at a Hackathon We are modernizing a laser marking system to use image recognition, we created an app when it was in alpha for a client to detect their hinges Backing up data and storage which is very necessary

What do you like best about Google Cloud AutoML?

You can use Google’s human labeling service to have real-life people annotate or clean your labels to make sure your models are being trained on high-quality data. It helps to make sure your data is clean and your organization can run smoother. It has a simple interface.

What do you dislike about Google Cloud AutoML?

it has an amazing UI and provides a great interface but it is not the cheapest option out there. Especially if you need custom training the price can be pretty hefty.

Recommendations to others considering Google Cloud AutoML:

They provide integration with human labeling which is great for smaller companies to have access to a much larger data base and technology.

What problems is Google Cloud AutoML solving and how is that benefiting you?

Cloud AutoML is fully integrated with other Google Cloud services, providing a method of access across the entire Google Cloud service line. We are able training data in Cloud Storage. To generate a prediction on your trained model, simply use the existing Vision API by adding a parameter for your custom model or use Cloud ML Engine's online prediction service.

What do you like best about Google Cloud AutoML?

The tool has facilitate the creation of machine learning models that we use in other applications.

What do you dislike about Google Cloud AutoML?

There is still a lot of art associated with training machine learning models. Some scaffolding to help with how to tune models would be helpful.

What problems is Google Cloud AutoML solving and how is that benefiting you?

We improved the accuracy of our machine learning in an important application. We were shooting for 99.9% accuracy. Our raw ML model got us to 90%, and with some pre- and post-processing of the data, we were able to get close to 99%.