Data Science Virtual Machine
Data Science Virtual Machine (DSVM) is a cloud server with an out-of-the-box operating system image and a pre-installed tools For machine learning developers and data analysts.
With DSVM you can:
- develop applications for chatbots, recommendation services, recognizing objects in photos and videos, speech synthesis and recognition, and prediction services;
- to train the models;
- perform experiments with data.
Tools
Tools are pre-installed in the image for DSVM (more details in their official documentation):
- Python 3.10;
- pip;
- PyTorch;
- TensorFlow;
- JupyterLab;
- Jupyter Notebook;
- Keras;
- scikit-learn;
- NumPy;
- SciPy;
- pandas;
- NLTK;
- OpenCV;
- CatBoost;
- XGBoost;
- LightGBM.
Create DSVM
Use the instructions Create a cloud server.
Select:
- source is a ready-made Data Science VM image (Ubuntu 22.04 LTS 64-bit);
- configuration — with RAM from 2 GB and boot disk capacity from 40 GB. For fast model training we recommend using fixed line configurations GPU Line or arbitrary configurations with GPU — they allow up to eight GPUs to be used.
Start JupyterLab
-
Open a page in your browser
http://<ip_address>Specify
<ip_address>— The public IP address of the cloud server can be viewed in control panels: from the top menu, press Products → Cloud servers → server page → tab Ports → column Public IP. -
Enter the password. By default, the UUID of the cloud server is used as the password. You can view it in control panels: from the top menu, press Products → Cloud servers → server page → alphanumeric value under the server name.