The platform also plans to offer the ability to rent GPUs from their network, allowing users to have complete control. Users can deploy their models for inference, either trained or fine-tuned on the platform, and choose from batch prediction or live endpoint methods. The platform is also working on the ability for users to publish their models for others to use, providing an earning opportunity. Users can join the platform's Discord for updates, feedback, feature requests, and tech support.
Key takeaways:
- Users can train models for free using credits during the BETA phase and earn rewards by contributing their GPUs to the network.
- The platform offers distributed training for users' own models, with no hidden costs and only pay for the training time used.
- Support for Google Colab is available to ensure users' code runs smoothly, and a no-code solution for fine-tuning models is coming soon.
- Users can deploy their models for inference, choose from batch prediction or live endpoint methods, and have the potential to earn by publishing their models for others to use.