- 谁
- AWS
- 引文
In this post, we walk through one complete example on a single SageMaker HyperPod cluster.
发生了什么
AWS 演示了如何使用 SageMaker HyperPod Data Capture 记录推理流量、通过 Athena 分析数据、创建 SQL 标注策略，并在同一集群上微调支持请求路由模型。在这项合成数据演练中，经过 60 秒的 LoRA 微调，模型在 280 个留出工单上的准确率从 83.9% 提升至 100%；另有一组 24 个手写探测样本，得分为 20/24。文章在 GitHub 仓库中提供了脚本、清单、SQL、数据集和预期输出。
推荐理由
这项演练展示了捕获的推理数据如何经过 Athena 分析、SQL 标注、微调和评估，并包含在留出工单和手写工单上的结果。
接下来读什么
查看这件事的完整进展AWS Demonstrates a SageMaker HyperPod Workflow for Capturing Data and Fine-Tuning Models这件事到目前为止
AWS demonstrated a SageMaker HyperPod Data Capture workflow that records inference traffic, analyzes it with Athena, applies a SQL labeling policy, and fine-tunes a support router on the same cluster. Data Capture is available on SageMaker HyperPod clusters with the inference operator. In the synthetic walkthrough, a 60-second LoRA fine-tune raised held-out accuracy from 83.9% to 100% across 280 tickets; a separate set of 24 hand-written probes scored 20 out of 24. AWS provided scripts, manifests, SQL, datasets, and expected outputs in a GitHub repository.