Wolfram Research announced the release of the new version of your programming language Wolfram Language and Wolfram Mathematica 12.1. The number of new features detailed in a blog post by creator Stephen Wolfram is quite extensive and has a lot to offer.
In this new version 12.1 the incorporation of Julia and R stands out to the collection of external languages, which means that the capabilities of the system should now be more accessible.
Besides it is mentioned that users who regularly use the neural network repository tungsten they will find 25 new types of networks, including the popular BERT language rendering model and the generative preformed transformer 2 used for text generation systems.
The system now it also comes with a token NetGANOperator and a TrainingUpdateSchedule option, which are intended, for example, to allow the general NetTrain functions in Wolfram to work with conflicting generative networks such as those often used in unsupervised learning or empowerment
“In version 12.1, we added Julia, Ruby, and R to our collection of external languages. Of course, there are all kinds of practical problems. We need to ensure that there is a proper installation on a user's computer and that the data types used in the programs can be meaningfully converted to the Wolfram Language
"It is actually very practical ... For example, you can create a function in the external language, which is then symbolically represented in the Wolfram Language as an ExternalFunction object and, when called, executes the code in the external language"
Other than that, the import of new neural network implementations should be a bit easier in the future, as version 12.1 now supports ONNX, an open format for representing machine learning models. Those working in image processing get more help with additions like FindImageText, which detects text in an image and marks it, while audiophiles will take advantage of SpeechInterpreter and SpeechCases.
Wolfram's team also modified data sets. Users will be delighted to know that now can set default values for number of rows and columns that will be displayed and can now better control the appearance of a dataset.
Data beyond what can be viewed on a laptop is stored directly on the laptop in the new version, which means that it will always be accessible once it is reopened. Meanwhile, two-dimensional data can now be captured and viewed using the experimental TableView function.
With version 12.1, packages have become a fully supported part of the Wolfram Language.
Wolfram himself described them in his introduction as modular packages of code and other resources to provide functionality.
“A paclet can configure the code to run at boot time. You can define symbols whose definitions will be loaded automatically. He can install the documentation. You can put items on menus. And in general, you can configure resources for use in almost any part of the rather complex structure of an implemented Wolfram Language system.
The structure of the paclet file contains "assets or resources of various types, as well as a special file PacletInfo.wl that defines how the paclet should be integrated into a Wolfram Language system."
And since the packages are usually distributed in a single compressed file, they are easy to share, so a central point of contact to do so is already in the works and should be available soon.
However, while the function repository preserves the consistency and structure of the language, the Paclet repository will not necessarily do so and is seen as a way to share entire environments for specific tasks.
This is just part of all the changes that are included in this new version, if you want to know the complete list you can consult the details in the following link.