add NVFP4 formal document

[xin3he]

User description

Type of Change

documentation

---

PR Type

Documentation, Enhancement

---

Description

• Added NVFP4 quantization documentation

• Updated README.md to include NVFP4 link

---

Diagram Walkthrough

flowchart LR
  A["Update README.md"] -- "Add NVFP4 link" --> B["Add PT_NVFP4Quant.md"]

Loading

File Walkthrough

Relevant files

Documentation

README.md Update README.md for NVFP4                                                              README.md Updated table structure to accommodate NVFP4 link Added new row for NVFP4 Quantization +11/-6    PT_NVFP4Quant.md Add NVFP4 Quantization Documentation                                          docs/source/3x/PT_NVFP4Quant.md Added comprehensive documentation for NVFP4 Quantization Included introduction, API usage, examples, and references +83/-0

---

[PRAgent4INC]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Formatting The table formatting might not render correctly in all Markdown viewers. Consider using a more universally compatible format or adding a note about viewer compatibility. <table> <tr> <th>Format Name</th> <th>Element Data type</th> <th>Element Bits</th> <th>Scaling Block Size</th> <th>Scale Data Type</th> <th>Scale Bits</th> <th>Global Scale Data Type</th> <th>Global Scale Bits</th> </tr> <tr> <td>NVFP4</td> <td>E2M1</td> <td>4</td> <td>16</td> <td>E4M3</td> <td>8</td> <td>FP32</td> <td>32</td> </tr> </table>

[PRAgent4INC]

PR Code Suggestions ✨

[xin3he]

Targeting 3.7, will change example usage to AutoRoundConfig and update this doc again.