OpenTSLM launches on SageMaker JumpStart
Understand this story
OpenTSLM, an open-source time-series language model developed by Aionic Labs, is now available on Amazon SageMaker JumpStart, and users can deploy it in their own AWS environments. The model combines numerical measurements such as device sensor data with textual context such as maintenance records and market news to output status summaries, recommended actions, and expected outcomes. Nexthink has already used it for its personal IT agent Spark, identifying potential issues from device telemetry such as CPU, memory, and disk, and guiding employees through fixes. On deployment requirements, OpenTSLM needs a single GPU with at least 24 GB of VRAM that supports bfloat16, with 32 GB of host memory recommended; the default instance type is ml.g6.2xlarge (one 24 GB NVIDIA L4 GPU, 32 GB host memory). Users need to check the quota for their chosen instance in the Service Quotas console. Deployed endpoints incur ongoing charges until deleted.
AI overview based on source reports · Updated 6 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- AWS Machine Learning BlogReason over time-series data with OpenTSLM on Amazon SageMaker JumpStart
Aionic Labs' open-source OpenTSLM family of time-series language models, which combine numerical measurements with text context, is now available in Amazon SageMaker JumpStart for deployment to SageMaker AI real-time inference endpoints. The post describes Nexthink's use of TSLMs in its Spark IT agent, walks through deploying OpenTSLM-Flamingo (Gemma 3 4B plus a time-series encoder, 5.24B parameters) via SageMaker Studio or the SageMaker Python SDK, and shows a fine-tuning tutorial using ARFBench data loaded through TimeNet, where accuracy on 135 held-out questions rose from 5.0% to 45.9%.
Activity for this story
- Current comparable activity
- 8
- Peak comparable activity
- 10Nov 18 03:00
- Change in the past 24 hours
- –
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.