Support single or multi dataset path for training

[mingxueg-nv]

Fixes #

Changes proposed in this pull request:

• Support single or multi dataset path for training

torchrun --nproc_per_node=$NUM_GPUS --master_port=29500 \
    gr00t/experiment/launch_finetune.py \
    --base_model_path  nvidia/GR00T-N1.6-3B \
    --dataset_path /dataset1 /dataset2 /dataset3 ...  \
    --embodiment_tag NEW_EMBODIMENT \
     ....

Before submitting

• I've read and followed all steps in the Making a pull request
  section of the CONTRIBUTING docs.
• I've updated or added any relevant docstrings.
• If this PR fixes a bug, I've added a test that will fail without my fix.
• If this PR adds a new feature, I've added tests that sufficiently cover my new functionality.