Feature support pipeline name in base_job_name at ModelTrainer

**Describe the feature you'd like**
i want to store the all files for jobs in a pattern like pipeline_name/job_name at this moment this is not possible for ModelTrainer class because train() function will call _create_training_job_args function which creates a s3 path like this:
s3://mybucket/my_pipeline_name/my_job_name/my-pipeline-name/my-job-name-20260318100943/input/sm_drivers/
or if the length of the base_job_name is longer then 63 the path could look like this:
s3://mybucket/my_pipeline_name/my_job_name/my-pipeline-name-extra-long-name/my-job-/input/sm_drivers/

**How would this feature be used? Please describe.**
This feature should be designed to help to store ModelTrainer files more consistently.
One suggestion for an S3 path would be {self.base_job_name}/input/20260318103152/sm_drivers/

**Additional context**
The Following Screenshot shows which code needs to be modified.

从 ModelTrainer.train() 开始，跟踪它对 _create_training_job_args() 的调用，issue 显示 S3 路径就是在这里拼接的。检查 base_job_name 的长度是如何处理的，并找出相关测试或路径构建入口。当 ModelTrainer 按请求的一致模式存储 job 文件，且不会为过长的名称生成格式错误的路径时，即表示完成。

由索引模型根据 Issue 内容生成。