Get Started
Last updated
Last updated
Log In: Open the Kaisar GPU Container portal and log in with your credentials.
Dashboard Overview: Upon logging in, you will be directed to the dashboard, which provides an overview of your resources and recent activities.
Dashboard Metrics: Displays key statistics:
Total GPUs Created: 210 (with a 55% increase)
Current Running GPUs: 3,052 (with a 14% decrease)
Total Compute Hours Served: 2,300 (with a 5% increase)
Total Memory Used: 173,000 units (with an 8% increase)
Usage Graph: Shows GPU usage over time, with consumption and top-up trends.
GPU Pricing: Lists available GPU chips with pricing and quantity details.
Network Map: Visualizes the distribution of GPUs by country.
Left Navigation Bar: Provides navigation links to different sections, such as Dashboard, Templates, GPU Container, Bare Metal GPUs, K8S Cluster, Ray Cluster, Account Settings, and Transactions.
Use the left-hand menu to navigate between different services. The primary service available currently is the GPU Container(highlighted on left hand in blue box in screenshot below), with other services such as Elastic GPUs, Bare Metal GPUs, Ray Node, K8S Cluster, and Ray Cluster coming soon.
Click on "GPU Container" to view information about the list of deployed containers. See screenshot below.
The screenshot above contains the following data:
GPU Container List: Shows a list of active GPU containers with their status, model, compute hours remaining, and created date.
Deploy GPU Container: A button to deploy new GPU containers.
Search and Filters: Options to filter containers by model and status
3. Click on a specific container from the list to view detailed information about each GPU container.See screenshot below for reference.
The screenshot above contains the following data related to GPU container:
Container Usage: Displays detailed usage statistics for a specific container, including size usage, memory usage, block IO read/write bytes, and network received/transmitted bytes.
Container Logs: Shows logs for the container activity.
Container Information: Includes container ID, uptime, compute hours, and connectivity tier.
Action Buttons: Options to terminate, extend, or connect to the container.
Click on "Templates" on the left hand to browse pre-built templates that can be used for GPU container deployment. Categories include GPU, Machine Learning, and Linux.
List of various templates available for GPU, Machine Learning, and Linux environments are mentioned below with their use cases.
Tool | Description |
---|---|
PyTorch | An open-source deep learning framework for research and development of new models, training and deployment of neural networks for image recognition, NLP, and generative models, and developing custom loss functions and optimizers. |
TensorFlow | An end-to-end open-source platform for building and deploying machine learning models for production-grade applications, large-scale distributed training and hyperparameter tuning, and implementing deep learning models for object detection, speech recognition, and text classification. |
Jupyter Notebook | An open-source web application for interactive data analysis and visualization, prototyping and experimenting with machine learning models, and teaching programming, data science, and machine learning concepts. |
Stable Diffusion | A generative model for generating images for creative projects, advertising, and content creation, developing applications that convert textual descriptions into visual representations, and research in GANs and image synthesis techniques. |
Visual Studio Code | A lightweight but powerful source code editor that supports a wide range of programming languages and frameworks, featuring debugging, embedded Git control, syntax highlighting, and code completion. |
Nvidia Cuda | A parallel computing platform and API model created by NVIDIA for using NVIDIA GPUs for general-purpose processing, providing significant performance improvements in computational tasks. |
Choose a template suitable for your needs, such as PyTorch. Selecting a particular template will provide details about the PyTorch template, including setup instructions, tag information, and components (like PyTorch version, CUDA, cuDNN, etc.).
Environment Variables: Variables needed to deploy the container, such as public key and Jupyter token.
Exposed Ports: Specifies ports for SSH and HTTP access.
Deploy Button: To deploy the container with the selected template.
See screenshot below for reference.
After selecting a template click on +Deploy button to initiate the deployment process. The button is highlighted in screenshot below on the upper right for reference.
Manage Account and Transactions
Manage Account Details
From the dashboard, locate the navigation panel on the left side of the screen and click on the Settings option under the Account section. See screenshot below for reference.
The screenshot above contains the Account section where you can update the following details:
Customer ID: This is auto-generated and not editable.
Full Name: Enter your full name in the provided field.
Email: Ensure your email is correct; you can update it if necessary.
Password: Click on Change Password to update your password. Enter your new password and confirm it.
Click on Save Change to apply any updates.
Managing API Keys
In the API Keys section, click on the + API Key button.
Follow the prompts to generate a new API key.
Ensure you store this key securely, as it grants access to your Kaisar Cloud account.
Managing SSH Public Keys
In the SSH Public Keys section, click on the SSH Public Key button.
Enter your public key in the provided field.
Click on Update Public Key to save the changes.
This allows for uninterrupted terminal access using the associated private key.
From the navigation panel, click on Notifications under the ACCOUNT section to view all notifications related to your account. See screenshot below for reference.
Notification Type: Use the dropdown menu to filter notifications by type and click on Mark all as read to mark all notifications as read. See screenshot below for reference.
Managing Transactions
From the dashboard, locate the navigation panel on the left side of the screen and click on the Transactions page.
The screenshot above displays the following data:
In Summary section, you can view your total Top Up, Payout, and Refund amounts
Transactions History Section: Here you can view all transaction records, including type, amount, status, and date.
Filters: To refine your search Transaction Type, Status, and Choose Date filters can be used.