Frequently Asked Questions
Q: Where do you source your GPUs from?
We partner with a diverse network of cloud providers, both large and small, to offer a wide range of GPUs with competitive pricing and high availability. We continuously evaluate and onboard new providers to ensure we can deliver the best possible service and selection.
This dynamic approach allows us to:
- Optimize for cost-effectiveness: By leveraging multiple providers, we can identify the most cost-efficient options for your needs
- Maximize availability: Our diverse network helps mitigate the risk of GPU shortages
- Offer cutting-edge hardware: We prioritize providers who offer the latest and most powerful GPUs
Pricing and availability may fluctuate over time. We strive to provide transparent and up-to-date information on our pricing page and through the CLI.
Q: What machine learning frameworks do you support?
Trainwave supports all major machine learning frameworks, including TensorFlow, PyTorch, Hugging Face Transformers, PyTorch Lightning, and XGBoost. We provide pre-built Docker images with these frameworks already installed. You can also use your own custom Docker images for specific requirements.
See Images for the full list of available images.
Q: How do I get support?
- Email: support@trainwave.ai
- In-app chat: Use the chat widget in the Trainwave web UI to connect with our support team in real-time
- Community: Join our Discord for community discussions
Q: What security measures do you have in place?
Security is a top priority at Trainwave. We employ a range of measures to protect your data and ensure the integrity of your training jobs:
- Isolated environments: Your jobs run in secure, isolated Docker containers
- Secure secrets management: A dedicated system for storing and managing sensitive information
- Compliance certifications: Trainwave is committed to meeting industry security standards
Q: What are your plans for the future?
We are continuously developing new features and enhancements. Our roadmap includes:
- Enhanced project management: Spending controls, project-level metrics, and granular access control
- Simplified model deployment: Streamlined workflows for deploying trained models
- Advanced collaboration features: Improved tools for team collaboration and project sharing
We’re excited to keep building the best platform for AI/ML training in the cloud!