We are pleased to announce that today we released Practicus AI v22.11 with many new and exciting features.
Some of our new features in this release include:
- Local container support: In addition to Practicus AI App, you can now run all the advanced cloud features locally on your computer and fully offline. This will help you reduce your cloud bill. Please read below to learn more!
- Improved GPU execution engine: We optimized our RAPIDS implementation to be able to handle larger volumes of data.
- Improved Apache Spark implementation: Our platform executes native with Pandas, DASK and RAPIDS (single or distributed). That being said, we are also committed to making Spark a first-class citizen since day one. With this release, we further improved our Spark engine implementation efficiency.
- 40+ UI improvements for a better user experience
Worker Nodes in a local container
Local container support was the #1 requested Practicus AI feature for a while.
Before we jump into local container benefits and how to use them, here’s a quick reminder on Practicus AI platform elements.
High level overview of Practicus AI platform elements
Some of the local container benefits
- Runs 100% private and securely on your computer.
- Lowest cost option and has forever free tier.
- No limitations on CPU/RAM capacity for the Free Tier. (Cloud option limits to 2 vCPUs and 1 GM RAM)
- Most new generation laptops have high compute capacity. You can easily process 50 million+ row datasets and fairly complex AutoML problems.
- You can switch back and forth to the cloud. E.g. You can use your computer for most use cases, switch to using the cloud with one click for very complex problems requiring GPUs (assuming your computer doesn’t have GPUs), and then switch back to continue on your computer when done.
- If you have a Practicus AI Enterprise license, all professional features are unlocked, and you get the same experience as the pro cloud option.
- Fairly simple installation and ready in 5-10 minutes.
- No credit card is needed to create a cloud account.
Setup local container support in 5-10 minutes
1) Install Docker Desktop or Podman Desktop on your computer
In order to run a container on your computer, you need to first install a container engine.
Docker is the most popular option: Install Docker Desktop
Although Docker Desktop is free, there has been some licensing changes in the recent years.
Podman is a great Docker alternative: Install Podman Desktop
Once the installation is completed, simply run Docker or Podman Desktop and confirm the container engine is running.
Active Docker Desktop View
Active Podman Desktop View
2) Pull (download) Practicus AI container image to your computer
Additional Practicus AI software is bundled inside a container image. You need to pull this package on your computer before using it.
- Open Practicus AI App settings (preferences in macOS) dialog and navigate to the Container section.
- If you have a Practicus AI Enterprise license, enter your email to activate and unlock all features. If not, you can use the free tier. Please note that Professional pay-as-you-go license option is not available for local containers. Compare license options.
- Choose a container engine, Docker or Podman, and confirm in the app that the engine is running.
- Click the Pull (download) Practicus AI container image button
- A command prompt window (terminal in macOS) will open to start the pull. This one-time download task can take anywhere between 5 – 20 minutes, depending on your internet speed.
- Once the container pull is completed, go back to the app and click refresh to view active Practicus AI images on your computer. Confirm you successfully pulled the container image.
- Click New Worker Node button to open Worker Nodes tab.
- Click Save to close settings.
- In the Worker Nodes tab, select local container as Cloud Region.
- Click Launch New button to start a Worker Node.
- When navigating cloud data sources in the Explore tab, you can switch between local and cloud Worker Nodes by using the drop-down list at the top right.
- Practicus AI app also attaches (mounts) container_shared folder, so you can easily copy files back and forth between your file system and the container. Simply open Windows Explorer (Finder in macOS), navigate to: your home folder / practicus / container_shared and copy files. Then navigate to Worker Node Files in Explore tab, and the files you copied will be visible under container_shared folder. Click Reload button at the top if you recently copied files.