Contribute example with Sagemaker SDK v3

**Describe the feature you'd like**
Recently, the SDK has been upgraded to V3, and it still lack of hand-on example with a customized algorithm container (scikit-learn, jax,...) along with old dependencies ([sagemaker training toolkit](https://github.com/aws/sagemaker-training-toolkit/tree/master)). I've worked on a company project on using Sagemaker SDK v3 recently, and want to contribute my implementation as an example to Sagemaker SDK documentation.

**How would this feature be used? Please describe.**
This example would enhance user experience with Sagemaker SDK v3

**Describe alternatives you've considered**
The example will use 2 principles,
- Bring your own container, customize the algorithm
- Pipeline declaration(runtime, control logic) and dataflow within pipeline (Property file, S3)

该 issue 未指定任何 repository 文件、测试或示例入口点；先定位 SageMaker SDK v3 文档示例，并检查链接的 sagemaker-training-toolkit。明确自定义算法容器和 pipeline 数据流的范围，然后验证该示例涵盖自带容器的使用，并能按文档所述运行。

由索引模型根据 Issue 内容生成。