Merge pull request #284 from NYU-RTS/Amanda-dong-patch-1

update intro

Author: Amanda-dong

docs/hpc/08_ml_ai_hpc/01_intro.md

@@ -4,4 +4,7 @@ Many people are interested in running ML and/or AI workflows on the HPC resource
 
 For ML we'll cover two prominent open-source deep learning frameworks, PyTorch and TensorFlow.  We'll show how to start with a single GPU example and then show more sophisticated examples. 
 
-For AI we'll cover how to run a Hugging Face Large Language Model (LLM) and how to fine tune an LLM.
+For AI we'll cover how to run a Hugging Face Large Language Model (LLM) and how to fine tune an LLM. Specifically, we provide guides on:
+1. LLM Inference: Use the standard Hugging Face transformers library for basic tasks, and also introduce vLLM, a high-throughput serving engine that offers faster inference and an OpenAI-compatible API.
+2. vLLM CLI: We provide examples of using the vllm command-line tool to quickly serve models and engage in interactive chat sessions.
+3. Fine-tuning LLMs: We provide a practical example of fine-tuning the Gemma model to follow specific instructions. This section compares the original model with our improved version, showing how to achieve better response quality on Torch.