PostgreSQL (alternately Postgres) is a free and open source object-relational database system boasting over 30 years of active development, reliability, feature robustness, and performance. It supports SQL and is designed to support various workloads flexibly.
N/A
TensorFlow
Score 7.7 out of 10
N/A
TensorFlow is an open-source machine learning software library for numerical computation using data flow graphs. It was originally developed by Google.
PostgreSQL is best used for structured data, and best when following relational database design principles. I would not use PostgreSQL for large unstructured data such as video, images, sound files, xml documents, web-pages, especially if these files have their own highly variable, internal structure.
TensorFlow is great for most deep learning purposes. This is especially true in two domains: 1. Computer vision: image classification, object detection and image generation via generative adversarial networks 2. Natural language processing: text classification and generation. The good community support often means that a lot of off-the-shelf models can be used to prove a concept or test an idea quickly. That, and Google's promotion of Colab means that ideas can be shared quite freely. Training, visualizing and debugging models is very easy in TensorFlow, compared to other platforms (especially the good old Caffe days). In terms of productionizing, it's a bit of a mixed bag. In our case, most of our feature building is performed via Apache Spark. This means having to convert Parquet (columnar optimized) files to a TensorFlow friendly format i.e., protobufs. The lack of good JVM bindings mean that our projects end up being a mix of Python and Scala. This makes it hard to reuse some of the tooling and support we wrote in Scala. This is where MXNet shines better (though its Scala API could do with more work).
Theano is perhaps a bit faster and eats up less memory than TensorFlow on a given GPU, perhaps due to element-wise ops. Tensorflow wins for multi-GPU and “compilation” time.
Postgresql is the best tool out there for relational data so I have to give it a high rating when it comes to analytics, data availability and consistency, so on and so forth. SQL is also a relatively consistent language so when it comes to building new tables and loading data in from the OLTP database, there are enough tools where we can perform ETL on a scalable basis.
The data queries are relatively quick for a small to medium sized table. With complex joins, and a wide and deep table however, the performance of the query has room for improvement.
There are several companies that you can contract for technical support, like EnterpriseDB or Percona, both first level in expertise and commitment to the software.
But we do not have contracts with them, we have done all the way from googling to forums, and never have a problem that we cannot resolve or pass around. And for dozens of projects and more than 15 years now.
Community support for TensorFlow is great. There's a huge community that truly loves the platform and there are many examples of development in TensorFlow. Often, when a new good technique is published, there will be a TensorFlow implementation not long after. This makes it quick to ally the latest techniques from academia straight to production-grade systems. Tooling around TensorFlow is also good. TensorBoard has been such a useful tool, I can't imagine how hard it would be to debug a deep neural network gone wrong without TensorBoard.
The online training is request based. Had there been recorded videos available online for potential users to benefit from, I could have rated it higher. The online documentation however is very helpful. The online documentation PDF is downloadable and allows users to pace their own learning. With examples and code snippets, the documentation is great starting point.
Although the competition between the different databases is increasingly aggressive in the sense that they provide many improvements, new functionalities, compatibility with complementary components or environments, in some cases it requires that it be followed within the same family of applications that performs the company that develops it and that is not all bad, but being able to adapt or configure different programs, applications or other environments developed by third parties apart is what gives PostgreSQL a certain advantage and this diversification in the components that can be joined with it, is the reason why it is a great option to choose.
Keras is built on top of TensorFlow, but it is much simpler to use and more Python style friendly, so if you don't want to focus on too many details or control and not focus on some advanced features, Keras is one of the best options, but as far as if you want to dig into more, for sure TensorFlow is the right choice
Easy to administer so our DevOps team has only ever used minimal time to setup, tune, and maintain.
Easy to interface with so our Engineering team has only ever used minimal time to query or modify the database. Getting the data is straightforward, what we do with it is the bigger concern.
