[MRG] [POC] using langchain to quickly POC training and eval steps

[HuaizhengZhang]

Closes #8

• understand stages
• data loading
• data engineering
• model selection (huggingface)
• model training (wandb)

future PRs

• model evaluation (fairness, accuracy)
• model deployment (streamlit)

Before submitting this PR, please make sure you have:

• confirmed all checks still pass OR confirm CI build passes.
• verified that any code or assets from external sources are properly credited in comments and/or in
  the credit file.

[HuaizhengZhang]

@huangyz0918 I will use langchain to quickly build several POCs and you can refactor them as you want

[huangyz0918]

@huangyz0918 I will use langchain to quickly build several POCs and you can refactor them as you want

Ok, but I will not use langchain to build the system. PS. please rebase since your OpenAI key are uploaded in the git

[huangyz0918]

I am working on using the openAI to load the data tonight, will refer to your poc code

[HuaizhengZhang]

@huangyz0918 merge first and will start to test and enhance your code