Resources and Usage
There are two types of resources with quota limits that users need when using HyperAI:
- Storage resources
- Computing resources
Below we will introduce these two types of resources in detail and explain how to view, purchase, manage, and clean up resources.
Storage Resources
Since container execution, data uploads, and log printing all generate data, and this data is permanently stored in HyperAI, the platform calculates a total size for all this data, representing the user's total storage resource usage. Each user will receive 20GB of free storage space after registration. You can view your current storage resource usage and total quota in "Resource Usage" - "Overview".
The following actions will consume additional storage space:
- Uploading data to "Data Warehouse"
- Starting a new container and generating new data in the container's "Workspace"
- Creating the working directory as a data warehouse version
The following actions will release storage space:
- Deleting a dataset or a specific dataset version (same applies to models and model versions)
- Deleting an entire container or deleting a specific execution
:::caution For projects that have been inactive for one full calendar month — meaning there have been no executions or container start records — HyperAI will permanently delete them.You can hover your mouse over the “trash can” icon on the “Model Training” page to view the specific cleanup time for each container. :::
Computing Resources
Running any container requires computing resources of the corresponding "computing power" type. This resource is measured in time units and begins counting when a container of the corresponding "computing power" type is marked as "Running". Computing resources have the following characteristics:
- Once consumed, they cannot be refunded
- Once a container is started, the "computing power" selection cannot be changed. If you need to change it, please stop running and restart
- Currently, computing resources support 4 billing methods: "Pay-as-you-go", "Daily Package", "Weekly Package", and "Monthly Package". For more details, refer to Resource Billing Methods
- If you choose pay-as-you-go, when your account balance is depleted, running containers will be forcibly shut down by the system and marked as "Shut Down Due to Arrears". Please ensure your account balance is sufficient. It is recommended to enable SMS notifications in settings to track container running status in real-time
:::caution Note Containers only start billing after being marked as "Running" until they are marked as "Stopped"/"Execution Completed"/"Execution Error". In other words, containers marked as "Preparing" do not consume computing resources. :::
Checking Current Quota
View current resource usage in "Resource Usage" - "Overview":
Computing resources are displayed separately according to different computing power types.
Purchasing Additional Resources
After completing "Balance Recharge" in the Finance Center, you can select appropriate computing power based on task requirements when creating a new model training step:
Purchasing Computing Resources
Currently, computing resources support 4 billing methods: "Pay-as-you-go", "Daily Package", "Weekly Package", and "Monthly Package". For more details, refer to Resource Billing Methods. Users can view current computing resource usage in "Resource Usage" - "Overview", which displays current usage and total quota.
Purchasing Storage Resources
Storage resources can be purchased by entering "Finance Center" from the left navigation - click "Subscribe to Storage Space".
Storage resources are billed by "Storage Limit" through subscription:
You can see that each purchasable storage resource type consists of different storage packages in "monthly" units. "20GB 2.00.
Storage resources automatically renew by default. Reminder emails will be sent to users 72 hours and 24 hours before the end of a monthly cycle. If not handled, the corresponding amount will be automatically deducted from the balance at expiration.
:::danger If the balance is insufficient, renewal will fail, and the storage resource subscription will be automatically canceled in the next cycle. After automatic cancellation, you will no longer have additional storage space, affecting normal service usage. :::
Upgrading or Downgrading Storage Resources
Storage space subscriptions can be adjusted by entering "Finance Center" - click "Upgrade/Downgrade" on the right to adjust storage space size.
In the "Upgrade/Downgrade" interface, you can select to change the current storage quota, upgrading or downgrading based on different storage sizes selected:
- Selecting larger storage space than current means "upgrade", taking effect immediately by paying the difference;
- Selecting smaller storage space than current means "downgrade", taking effect with the new subscription next month;
- If selecting "Free Space", it means canceling the current storage space, which will not renew next month;
:::info All storage resources are added on top of the free 20GB base resource, meaning after selecting 20GB of storage space, the total storage space becomes 40GB. :::
Querying Resource Update Records
Resource addition records can be viewed in "Resource Usage" - "Resource Change Details":
Querying Usage
Specific resource consumption records can be viewed in "Resource Usage" - "Usage Details":
- After a container is closed, two records appear representing "computing resource" usage and "storage resource" usage respectively
- When a container is deleted and its storage resources are released, the storage space field displays as a negative number
- After a data warehouse is successfully uploaded, a "storage resource" usage record is added
- When a data warehouse or a data warehouse version is deleted, a negative usage entry is added, indicating storage resources have been released
Learn More
To learn more about resources and usage, refer to the following documents:
- Key Concepts - Learn more about HyperAI's resource management and usage
- Resource Billing Methods - Learn about different resource billing methods
- Data Warehouse - Learn how to manage and reuse datasets
- FAQ - View more frequently asked questions about resources and usage