fine-tuning notebooks and scripts

[sciarrilli]

Issue #, if available:

Description of changes:

Added 3 notebooks for fine-tuning section of workshop. Also added scripts folder with fine-tuning and eval scripts.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[ravees]

@sciarrilli - Great content. See initial feedback below. I'll provide more as I review:

1. Can you use the same pattern as noted in the other notebooks (for example agents-intro notebook) for pip install !pip install --upgrade -r requirements.txt -q
2. For loading the SageMaker managed MLflow tracking server ARN the TRACKING_SERVER_ARN variable from the stored value.
3. Should we add instructions for the participants to get the hf_token especially for participants new to HF?
4. Is there a reason to check in the notebook cell output to the repo?
5. Can you move the variable you set into a separate cell where possible and add a comment/markdown note to explain to the reader. Like pytorch_image = '763104351884.dkr.ecr.us-east-1.amazonaws.com/pytorch-training:2.8.0-gpu-py312-cu129-ubuntu22.04-sagemaker'
6. Remove the hardcoded variables like s3 bucket URL, tracking server arn. 'MLFLOW_TRACKING_URI', 'arn:aws:sagemaker:us-east-1:198346569064:mlflow-tracking-server/vlm-finetuning-server', "adapter_path": 's3://sagemaker-us-east-1-198346569064/qwen3-06b-fine-tuned/'

[sciarrilli]

@ravees

1. done
2. I don't understand what you are asking here?
3. let me double check this. we might not need to set the hf_token
4. done
5. done
6. let me know the mlflow app name in the workshop studio and i will retrieve mlflow server arn. i need to rewrite the eval workshop to pull down the tar.gz from the last training job.