Thank you for using our quantum machine learning cloud system. We have been providing a paid cloud service for businesses, previously utilizing our own proprietary interface. We are pleased to announce the adoption of Kubeflow, making it easier to conduct flexible research and development in quantum computing and machine learning across various environments, including our unique cloud, on-premises, and major cloud platforms.
Kubeflow is a framework designed to manage the training and deployment of machine learning models. By extending its capabilities to quantum computing, we have enabled more flexible development opportunities.
In addition to CPUs, our company promotes the integration of development environments with machine learning and the use of GPUs through NVIDIA cuQuantum. By utilizing our quantum computing framework developed with PyTorch, we enable seamless execution of quantum computing and machine learning, offering numerous benefits.
With the adoption of Kubeflow as the interface, we now offer more flexible management of GPU usage and data and model management.
The basic usage involves:
- Selecting the type of GPU to use (e.g., H100, RTX6000ada, T4, etc.)
- Choosing the Notebook image of the framework to use (e.g., PyTorch, cuQuantum, TensorFlow, etc.)
- Selecting the volume for data storage (e.g., personal, group shared, etc.)
- Choosing other computational resources such as CPU and memory (e.g., amount of CPU, memory, etc.)
This allows for a free combination within the scope of the contract, integrating these computations freely with quantum computers via API.
Furthermore, we have transitioned part of our cloud services, previously built on AWS, to our proprietary cloud system. This shift offers more flexibility in machine management and the ability to provide various services. Networks can also be freely chosen, enabling more flexible quantum computing research and development.
Additionally, by employing multiple GPUs, we can significantly accelerate quantum computing and machine learning. Adopting NVIDIA's official cuQuantum enables multi-GPU and multi-node computations, requiring the introduction of a dedicated framework called cuQuantum Appliance. This expansion provides numerous business benefits.
Our dedicated cloud infrastructure management team is constantly improving the system and adding features in response to requests. Businesses interested in advancing their quantum machine learning research and development using our cloud services are encouraged to consider this opportunity.