update readme

Signed-off-by: jihyeonRyu <[email protected]>

Author: jihyeonRyu

tutorials/llm/llama-3/README.rst

@@ -23,3 +23,6 @@ This repository contains Jupyter Notebook tutorials using the NeMo Framework for
    * - `Llama3 LoRA Fine-Tuning and Supervised Fine-Tuning using NeMo2 <./nemo2-sft-peft>`_
      - `SQuAD <https://arxiv.org/abs/1606.05250>`_ for LoRA and `Databricks-dolly-15k <https://huggingface.co/datasets/databricks/databricks-dolly-15k>`_ for SFT
      - Perform LoRA PEFT and SFT on Llama 3 8B using NeMo 2.0
+   * - `Llama3 Domain Adaptive Pre-Training <./dapt>`_
+     - `Domain-Specific Data <https://github.com/NVIDIA/NeMo-Curator/tree/main/tutorials/dapt-curation>`_
+     - Perform Domain Adaptive Pre-Training on Llama 3 8B using NeMo 2.0

tutorials/llm/llama-3/dapt/README.md

@@ -1,6 +1,6 @@
 # Training Code for DAPT (Domain Adaptive Pre-Training)
 
-[ChipNeMo](https://arxiv.org/pdf/2311.00176) is a chip design domain adapted LLM. Instead of directly deploying off-theshelf commercial or open-source LLMs, the paper instead adopts the following domain adaptation techniques: domain-adaptive tokenization, domain adaptive continued pretraining, model alignment with domain-specific instructions, and domain adapted retrieval models. Specifically, LLama 2 foundation models are continually pre-trained with 20B plus tokens on domain-specific chip design data, including code, documents, etc., and then fine-tuned with instruction datasets from design data as well as external sources. Evaluations on the resultant domain-adapted ChipNeMo model demonstrate that domain-adaptive pretraining of language models, can lead to superior performance in domain related downstream tasks compared to their base LLaMA2 counterparts, without degradations in generic capabilities.
+[ChipNeMo](https://arxiv.org/pdf/2311.00176) is a chip design domain adapted LLM. Instead of directly deploying off-theshelf commercial or open-source LLMs, the paper instead adopts the following domain adaptation techniques: domain-adaptive tokenization, domain adaptive continued pretraining, model alignment with domain-specific instructions, and domain adapted retrieval models. 
 
 Here, we share a tutorial with best practices on training for DAPT (domain-adaptive pre-training).