AWS SageMaker AI training sessions now GA
What you need to know
AWS has made training sessions in Amazon SageMaker AI serverless model customization generally available. The training session APIs let users run custom training loops, such as forward_backward() GPU operations, without managing the underlying infrastructure. Usage is billed per token processed, so no GPU time is charged while work runs on the user's own machine, such as a reward function scoring a batch. A running session counts against the concurrent-session quota until stop() is called or it expires, so AWS says to stop it when finished, including on error paths. Supported base models and the AWS Regions where each is available are listed in the training session APIs documentation.
Generated by AI from the reporting · updated 3 hours ago
Timeline
Follow the reports to see every side of the story.
- AWS Machine Learning BlogAWS makes SageMaker AI training sessions generally available for custom model post-training
AWS launched general availability of training sessions in Amazon SageMaker AI serverless model customization.
- AWS Machine Learning BlogAmazon SageMaker AI adds training session APIs for custom LLM training loops
Amazon SageMaker AI expanded serverless model customization with training session APIs that turn GPU-bound operations like forward_backward().
Heat trend for this story
- Current
- 9
- Peak
- 10Oct 22 19:00
- 24h change
- –
The trend compares only sources observed continuously throughout, so its range may be narrower than the current heat total. Move the pointer or tap the chart to see heat by the hour; use the left and right arrow keys on a keyboard.