feat(examples): add two-node QLoRA recipe

[HuiyingLi]

Summary

• Add a Llama 3.1 8B Instruct SQuAD QLoRA recipe for a 2-node, 16-GPU FSDP2 run.
• Configure 4-bit NF4 quantization, LoRA adapters, validation every 100 steps, and a 100-step run.

Validation

• Parsed the YAML config with PyYAML.
• Ran git diff --check on the new recipe.
• Previously validated the recipe shape with a 2-node Slurm run for 100 steps: job 12099765 completed and logged to W&B: https://wandb.ai/Nemo-automodel/huiyingl_workspace/runs/722ux2w9

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[HuiyingLi]

/ok to test 5414283