[Epic] Add data checkpointing and recovery

[markmc]

In order to ensure that progress is not completely lost if there is an error during a long-run data generation process, add a checkpointing system that saves intermediate states during data generation. This allows for more granular progress tracking and recovery.

See aakankshaduggal#19

• Add capability to library -- Add data checkpointing capability #222
• Enable checkpointing in the CLI -- Adding checkpoint dir when batching enabled instructlab#1915
• Enable batching in CLI (checkpointing is coupled with batching) Enable SDG batching with vLLM instructlab#1797
• File an issue to consider making use of save_freq in future -- checkpointing: consider allowing users to specify save frequency #225