add mxfp8 qat example.

[lkk12014402]

Type of Change

add mxfp8 qat example

Description

mxfp8 qat

[Copilot]

Pull Request Overview

This PR adds a new example for MXFP8 Quantization Aware Training (QAT) for large language models using Hugging Face transformers. The example demonstrates how to perform QAT on LLMs with the Neural Compressor library.

• Provides a complete QAT workflow with utilities for data processing and model quantization
• Implements training pipeline with MXFP8 quantization scheme integration
• Includes comprehensive documentation and configuration files for distributed training

Reviewed Changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
utils.py	Data processing utilities for the Daring-Anteater dataset and metrics calculation
requirements.txt	Dependencies for the QAT example including neural-compressor and transformers
main.py	Main training script with quantization integration and model export functionality
accelerate_config/fsdp1.yaml	FSDP configuration for distributed training
accelerate_config/ddp.yaml	DDP configuration for multi-GPU training
README.md	Documentation with step-by-step QAT workflow instructions

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}