Use llm export for refit

[yfw]

What does this PR do ?

This allows us to use NeMo's llm export mapping and transforms when converting megatron params to HF params for refit. The mapping and transforms in nemo_rl/models/megatron/converters/qwen2.py is currently copied from https://github.com/NVIDIA/NeMo/blob/c882b4885349b5a750147d242338ebcbc7058ef6/nemo/collections/llm/gpt/model/qwen2.py#L316-L358

Issues

List issues that this PR closes (syntax):

Usage

• You can potentially add a usage example below

# Add a code snippet demonstrating how to use this 

Before your PR is "Ready for review"

Pre checks:

• Make sure you read and followed Contributor guidelines
• Did you write any new necessary tests?
• Did you run the unit tests and functional tests locally? Visit our Testing Guide for how to run tests
• Did you add or update any necessary documentation? Visit our Document Development Guide for how to write, build and test the docs.

Additional Information

• ...