With the recent major release of PyTorch 2.0, new trainML job environments are available.
Use Large Language Model Checkpoints for Free
Start building your own ChatGPT-like applications with popular open source Large Language Models like GPT-J, GPT-NeoX, and BloomZ.
Attach Checkpoints to Edge Inference Devices
CloudBender™ Device Configurations have been expanded to allow attaching public and private Checkpoints to edge inference devices.
Create Checkpoints and Datasets from Job Outputs
Checkpoints and Datasets are now supported output destinations for trainML Training and Inference jobs.
Automatically Integrate Huggingface Models
Checkpoints can now be created directly from public or private Hugging Face models.
Free Public Checkpoints
Get started building models even faster by using Public Checkpoints.
Reusable Checkpoints Make Models Even More Flexible
trainML now supports the creation and use of Checkpoints to store immutable versions of large model weight files.
Google Drive Storage Source
Deliver Analytics To Customer Securely With Federated Inference
Analytics providers can now run their models directly on the trainML deployments of their customers. This allows the analytics provider to maintain and protect their intellectual property while providing analytics services inside their customers' secure, private infrastructure.
Take Inference to the Edge with GPU-enabled CloudBender Devices
Run real-time inference workloads on NVIDIA Jetson fully managed by CloudBender™.