- Create an instance
- Download demos
- Install third-party libraries
- Authorize RAM user accounts
- Mount NAS file systems
Data Science Workshop (DSW) is a deep learning development environment provided on the cloud for algorithm developers to write and run code. DSW uses the TensorFlow framework that is deeply optimized by the Machine Learning Platform for AI team. You can open the DSW terminal to install third-party libraries.
The following figure shows the user interface of DSW:
The user interface contains the files section on the left side, the code editing section in the middle, and the resource search section on the right side.
- Log on to Machine Learning Platform for AI and select DSW in the left-side navigation pane.
Select a region. Currently, only the China (Shanghai) and China (Beijing) regions are supported. The underlying GPU servers used by DSW and their prices vary depending on the region. For more information, see DSW pricing
Click Create Instance.
Enter the instance name, and select a resource type. For more information, see DSW pricing.
- Resource type descriptions
The system provides 5 GB of free storage space for each created instance. If you want to mount an NAS file system to the instance, purchase an NAS file system and enter the NAS file system ID into the field as shown in the following figure. You must click “Authorize” to authorize the NAS file system before you bind the NAS file system to the instance.
- Open the created instance to log on to DSW.
For new users, we recommend that you use the built-in demos to get familiar with DSW. In the left-side navigation pane, click Demos, select the demo that you want to download, and click Download.
You can view and use the downloaded demo in the File Browser.
Third-party libraries are required for code development in a Python environment. Machine Learning Platform for AI allows you to install third-party libraries. The installed third-party libraries are saved to the environment for further usage.
To install third-party libraries, open the terminal.
Install third-party libraries for Python 3:
pip install --user xxx
Install third-party libraries for Python 2:
source activate python2
pip install --user xxx
If you want to use a RAM user account to create a DSW instance and use it, you must first log on to Resource Access Management (RAM) to authorize the RAM user account.
Log on to the RAM console, click Policies in the left-side navigation pane, click the Custom Policy tab, and select Create Authorization Policy.
Select Blank Template, enter a policy name, and add the following content to the policy:
Click Users in the left-side navigation pane.
Authorize the RAM user account.
You must make sure that the NAS file system is purchased in the same region as DSW. To purchase NAS file systems, click https://www.aliyun.com/product/nas
Log on to the NAS console and follow the steps shown in the following figure:
- Create an NAS file system and set mount points.
- When you create the NAS file system, do not set the protocol type to NASPlus. Set the protocol type to NFS.
- Add a mount point and set the mount point type. ( If you want to download large files, set the network type to VPC. For more information, see Upload and download data. Otherwise, set the network type to Classic.If you set the network type to Classic, you must create a permission group and apply the permission group to the NAS file system. To create a permission group, click https://nas.console.aliyun.com/?spm=5176.6660585.774526198.1.2a316bf8NlExu3#/accessGroup/list
- Authorize the NAS file system
Skip this step if your NAS file system is already authorized.
Click the following link to go to the authorization page and then click Confirm Authorization Policy: NAS file system read permission authorization
Record the NAS file system ID. This ID is used to mount the NAS file system to the DSW instance.