torch-rechub 0.1.0__tar.gz → 0.3.0__tar.gz

{torch_rechub-0.1.0 → torch_rechub-0.3.0}/CHANGELOG.md

@@ -7,6 +7,47 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ---
 
+## [0.3.0] - 2026-02-05
+
+<!-- Release notes generated using configuration in .github/release.yml at main -->
+
+## What's Changed
+### ✨ 新特性 / Features
+* Inbatchsample by @zerolovesea in https://github.com/datawhalechina/torch-rechub/pull/128
+### 📝 文档更新 / Documentation
+* Enhance docs: ONNX, serving, tools, tutorials by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/165
+* Expand docs: community, guides, models, data by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/166
+### 🔄 其他变更 / Other Changes
+* Update favicon and logo images by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/164
+* Update docs imports to torch_rechub by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/167
+
+
+**Full Changelog**: https://github.com/datawhalechina/torch-rechub/compare/v0.2.0...v0.3.0
+
+---
+
+
+
+## [0.2.0] - 2026-01-11
+
+<!-- Release notes generated using configuration in .github/release.yml at main -->
+
+## What's Changed
+### 🐛 Bug 修复 / Bug Fixes
+* Update HSTULayer with L2 norm and SiLU attention by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/157
+* Add ONNX quantization scripts and tests, improve dataloader split  by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/161
+### 📝 文档更新 / Documentation
+* Update docs and README with new architecture image by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/154
+* Add comprehensive vector index docs (EN & ZH) by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/162
+* Update quick start and tracking docs in EN and ZH by @1985312383 in https://github.com/datawhalechina/torch-rechub/pull/163
+
+
+**Full Changelog**: https://github.com/datawhalechina/torch-rechub/compare/v0.1.0...v0.2.0
+
+---
+
+
+
 ## [0.1.0] - 2025-12-17
 
 <!-- Release notes generated using configuration in .github/release.yml at main -->

{torch_rechub-0.1.0 → torch_rechub-0.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: torch-rechub
-Version: 0.1.0
+Version: 0.3.0
 Summary: A Pytorch Toolbox for Recommendation Models, Easy-to-use and Easy-to-extend.
 Project-URL: Homepage, https://github.com/datawhalechina/torch-rechub
 Project-URL: Documentation, https://www.torch-rechub.com
@@ -31,7 +31,7 @@ Requires-Dist: transformers>=4.46.3
 Provides-Extra: annoy
 Requires-Dist: annoy>=1.17.2; extra == 'annoy'
 Provides-Extra: bigdata
-Requires-Dist: pyarrow~=21.0; extra == 'bigdata'
+Requires-Dist: pyarrow<23,>=21; extra == 'bigdata'
 Provides-Extra: dev
 Requires-Dist: bandit>=1.7.0; extra == 'dev'
 Requires-Dist: flake8>=3.8.0; extra == 'dev'
@@ -60,9 +60,13 @@ Requires-Dist: graphviz>=0.20; extra == 'visualization'
 Requires-Dist: torchview>=0.2.6; extra == 'visualization'
 Description-Content-Type: text/markdown
 
-# 🔥 Torch-RecHub - 轻量、高效、易用的 PyTorch 推荐系统框架
+<div align="center">
 
-> 🚀 **30+ 主流推荐模型** | 🎯 **开箱即用** | 📦 **一键部署 ONNX** | 🤖 **支持生成式推荐 (HSTU/HLLM)**
+![Torch-RecHub 横幅](docs/public/img/banner.png)
+
+# Torch-RecHub: 轻量、高效、易用的 PyTorch 推荐系统框架
+
+【⚠️ Alpha内测版本警告：此为早期内部构建版本，尚不完整且可能存在错误，欢迎大家提Issue反馈问题或建议。】
 
 [![许可证](https://img.shields.io/badge/license-MIT-blue?style=for-the-badge)](LICENSE)
 ![GitHub Repo stars](https://img.shields.io/github/stars/datawhalechina/torch-rechub?style=for-the-badge)
@@ -75,27 +79,21 @@ Description-Content-Type: text/markdown
 [![numpy 版本](https://img.shields.io/badge/numpy-1.19%2B-orange?style=for-the-badge)](https://numpy.org/)
 [![scikit-learn 版本](https://img.shields.io/badge/scikit_learn-0.23%2B-orange?style=for-the-badge)](https://scikit-learn.org/)
 [![torch-rechub 版本](https://img.shields.io/badge/torch_rechub-0.0.3%2B-orange?style=for-the-badge)](https://pypi.org/project/torch-rechub/)
+[![torchview](https://img.shields.io/badge/torchview-0.2%2B-green?style=for-the-badge)](https://github.com/mert-kurttutan/torchview)
 
 [English](README_en.md) | 简体中文
 
-**在线文档：** https://datawhalechina.github.io/torch-rechub/ （英文）| https://datawhalechina.github.io/torch-rechub/zh/ （简体中文）
+![架构图](docs/public/img/project_framework.png)
 
-**Torch-RecHub** —— **10 行代码实现工业级推荐系统**。30+ 主流模型开箱即用，支持一键 ONNX 部署，让你专注于业务而非工程。
+</div>
 
-![Torch-RecHub 横幅](docs/public/img/banner.png)
-
-## 🎯 为什么选择 Torch-RecHub？
+**在线文档：** https://datawhalechina.github.io/torch-rechub/zh/
 
-| 特性          | Torch-RecHub                | 其他框架   |
-| ------------- | --------------------------- | ---------- |
-| 代码行数      | **10行** 完成训练+评估+部署 | 100+ 行    |
-| 模型覆盖      | **30+** 主流模型            | 有限       |
-| 生成式推荐    | ✅ HSTU/HLLM (Meta 2024)     | ❌          |
-| ONNX 一键导出 | ✅ 内置支持                  | 需手动适配 |
-| 学习曲线      | 极低                        | 陡峭       |
+**Torch-RecHub** —— **10 行代码实现工业级推荐系统**。30+ 主流模型开箱即用，支持一键 ONNX 部署，让你专注于业务而非工程。
 
 ## ✨ 特性
 
+* **生成式推荐模型:** LLM时代下，可以复现部分生成式推荐模型
 * **模块化设计:** 易于添加新的模型、数据集和评估指标。
 * **基于 PyTorch:** 利用 PyTorch 的动态图和 GPU 加速能力。
 * **丰富的模型库:** 涵盖 **30+** 经典和前沿推荐算法（召回、排序、多任务、生成式推荐等）。
@@ -109,7 +107,6 @@ Description-Content-Type: text/markdown
 ## 📖 目录
 
 - [🔥 Torch-RecHub - 轻量、高效、易用的 PyTorch 推荐系统框架](#-torch-rechub---轻量高效易用的-pytorch-推荐系统框架)
-  - [🎯 为什么选择 Torch-RecHub？](#-为什么选择-torch-rechub)
   - [✨ 特性](#-特性)
   - [📖 目录](#-目录)
   - [🔧 安装](#-安装)
@@ -221,6 +218,8 @@ torch-rechub/             # 根目录
 
 本框架目前支持 **30+** 主流推荐模型：
 
+<details>
+
 ### 排序模型 (Ranking Models) - 13个
 
 | 模型          | 论文                                             | 简介                    |
@@ -236,7 +235,11 @@ torch-rechub/             # 根目录
 | **AutoInt**   | [CIKM 2019](https://arxiv.org/abs/1810.11921)    | 自动特征交互学习        |
 | **FiBiNET**   | [RecSys 2019](https://arxiv.org/abs/1905.09433)  | 特征重要性 + 双线性交互 |
 | **DeepFFM**   | [RecSys 2019](https://arxiv.org/abs/1611.00144)  | 场感知因子分解机        |
-| **EDCN**      | [KDD 2021](https://arxiv.org/abs/2106.03032)     | 增强型交叉网络          |
+| **EDCN**      | [KDD 2021](https://arxiv.org/abs/2106.03032)     | 增强型交叉网络  
+        |
+</details>
+
+<details>
 
 ### 召回模型 (Matching Models) - 12个
 
@@ -253,6 +256,10 @@ torch-rechub/             # 根目录
 | **STAMP**      | [KDD 2018](https://dl.acm.org/doi/10.1145/3219819.3219895)                     | 短期注意力记忆优先 |
 | **ComiRec**    | [KDD 2020](https://arxiv.org/abs/2005.09347)                                   | 可控多兴趣推荐     |
 
+</details>
+
+<details>
+
 ### 多任务模型 (Multi-Task Models) - 5个
 
 | 模型             | 论文                                                          | 简介               |
@@ -263,6 +270,10 @@ torch-rechub/             # 根目录
 | **AITM**         | [KDD 2021](https://arxiv.org/abs/2105.08489)                  | 自适应信息迁移     |
 | **SharedBottom** | -                                                             | 经典多任务共享底层 |
 
+</details>
+
+<details>
+
 ### 生成式推荐 (Generative Recommendation) - 2个
 
 | 模型     | 论文                                          | 简介                                         |
@@ -270,6 +281,8 @@ torch-rechub/             # 根目录
 | **HSTU** | [Meta 2024](https://arxiv.org/abs/2402.17152) | 层级序列转换单元，支撑 Meta 万亿参数推荐系统 |
 | **HLLM** | [2024](https://arxiv.org/abs/2409.12740)      | 层级大语言模型推荐，融合 LLM 语义理解能力    |
 
+</details>
+
 ## 📊 支持的数据集
 
 框架内置了对以下常见数据集格式的支持或提供了处理脚本：

{torch_rechub-0.1.0 → torch_rechub-0.3.0}/README.md

@@ -1,6 +1,10 @@
-# 🔥 Torch-RecHub - 轻量、高效、易用的 PyTorch 推荐系统框架
+<div align="center">
 
-> 🚀 **30+ 主流推荐模型** | 🎯 **开箱即用** | 📦 **一键部署 ONNX** | 🤖 **支持生成式推荐 (HSTU/HLLM)**
+![Torch-RecHub 横幅](docs/public/img/banner.png)
+
+# Torch-RecHub: 轻量、高效、易用的 PyTorch 推荐系统框架
+
+【⚠️ Alpha内测版本警告：此为早期内部构建版本，尚不完整且可能存在错误，欢迎大家提Issue反馈问题或建议。】
 
 [![许可证](https://img.shields.io/badge/license-MIT-blue?style=for-the-badge)](LICENSE)
 ![GitHub Repo stars](https://img.shields.io/github/stars/datawhalechina/torch-rechub?style=for-the-badge)
@@ -13,27 +17,21 @@
 [![numpy 版本](https://img.shields.io/badge/numpy-1.19%2B-orange?style=for-the-badge)](https://numpy.org/)
 [![scikit-learn 版本](https://img.shields.io/badge/scikit_learn-0.23%2B-orange?style=for-the-badge)](https://scikit-learn.org/)
 [![torch-rechub 版本](https://img.shields.io/badge/torch_rechub-0.0.3%2B-orange?style=for-the-badge)](https://pypi.org/project/torch-rechub/)
+[![torchview](https://img.shields.io/badge/torchview-0.2%2B-green?style=for-the-badge)](https://github.com/mert-kurttutan/torchview)
 
 [English](README_en.md) | 简体中文
 
-**在线文档：** https://datawhalechina.github.io/torch-rechub/ （英文）| https://datawhalechina.github.io/torch-rechub/zh/ （简体中文）
+![架构图](docs/public/img/project_framework.png)
 
-**Torch-RecHub** —— **10 行代码实现工业级推荐系统**。30+ 主流模型开箱即用，支持一键 ONNX 部署，让你专注于业务而非工程。
+</div>
 
-![Torch-RecHub 横幅](docs/public/img/banner.png)
-
-## 🎯 为什么选择 Torch-RecHub？
+**在线文档：** https://datawhalechina.github.io/torch-rechub/zh/
 
-| 特性          | Torch-RecHub                | 其他框架   |
-| ------------- | --------------------------- | ---------- |
-| 代码行数      | **10行** 完成训练+评估+部署 | 100+ 行    |
-| 模型覆盖      | **30+** 主流模型            | 有限       |
-| 生成式推荐    | ✅ HSTU/HLLM (Meta 2024)     | ❌          |
-| ONNX 一键导出 | ✅ 内置支持                  | 需手动适配 |
-| 学习曲线      | 极低                        | 陡峭       |
+**Torch-RecHub** —— **10 行代码实现工业级推荐系统**。30+ 主流模型开箱即用，支持一键 ONNX 部署，让你专注于业务而非工程。
 
 ## ✨ 特性
 
+* **生成式推荐模型:** LLM时代下，可以复现部分生成式推荐模型
 * **模块化设计:** 易于添加新的模型、数据集和评估指标。
 * **基于 PyTorch:** 利用 PyTorch 的动态图和 GPU 加速能力。
 * **丰富的模型库:** 涵盖 **30+** 经典和前沿推荐算法（召回、排序、多任务、生成式推荐等）。
@@ -47,7 +45,6 @@
 ## 📖 目录
 
 - [🔥 Torch-RecHub - 轻量、高效、易用的 PyTorch 推荐系统框架](#-torch-rechub---轻量高效易用的-pytorch-推荐系统框架)
-  - [🎯 为什么选择 Torch-RecHub？](#-为什么选择-torch-rechub)
   - [✨ 特性](#-特性)
   - [📖 目录](#-目录)
   - [🔧 安装](#-安装)
@@ -159,6 +156,8 @@ torch-rechub/             # 根目录
 
 本框架目前支持 **30+** 主流推荐模型：
 
+<details>
+
 ### 排序模型 (Ranking Models) - 13个
 
 | 模型          | 论文                                             | 简介                    |
@@ -174,7 +173,11 @@ torch-rechub/             # 根目录
 | **AutoInt**   | [CIKM 2019](https://arxiv.org/abs/1810.11921)    | 自动特征交互学习        |
 | **FiBiNET**   | [RecSys 2019](https://arxiv.org/abs/1905.09433)  | 特征重要性 + 双线性交互 |
 | **DeepFFM**   | [RecSys 2019](https://arxiv.org/abs/1611.00144)  | 场感知因子分解机        |
-| **EDCN**      | [KDD 2021](https://arxiv.org/abs/2106.03032)     | 增强型交叉网络          |
+| **EDCN**      | [KDD 2021](https://arxiv.org/abs/2106.03032)     | 增强型交叉网络  
+        |
+</details>
+
+<details>
 
 ### 召回模型 (Matching Models) - 12个
 
@@ -191,6 +194,10 @@ torch-rechub/             # 根目录
 | **STAMP**      | [KDD 2018](https://dl.acm.org/doi/10.1145/3219819.3219895)                     | 短期注意力记忆优先 |
 | **ComiRec**    | [KDD 2020](https://arxiv.org/abs/2005.09347)                                   | 可控多兴趣推荐     |
 
+</details>
+
+<details>
+
 ### 多任务模型 (Multi-Task Models) - 5个
 
 | 模型             | 论文                                                          | 简介               |
@@ -201,6 +208,10 @@ torch-rechub/             # 根目录
 | **AITM**         | [KDD 2021](https://arxiv.org/abs/2105.08489)                  | 自适应信息迁移     |
 | **SharedBottom** | -                                                             | 经典多任务共享底层 |
 
+</details>
+
+<details>
+
 ### 生成式推荐 (Generative Recommendation) - 2个
 
 | 模型     | 论文                                          | 简介                                         |
@@ -208,6 +219,8 @@ torch-rechub/             # 根目录
 | **HSTU** | [Meta 2024](https://arxiv.org/abs/2402.17152) | 层级序列转换单元，支撑 Meta 万亿参数推荐系统 |
 | **HLLM** | [2024](https://arxiv.org/abs/2409.12740)      | 层级大语言模型推荐，融合 LLM 语义理解能力    |
 
+</details>
+
 ## 📊 支持的数据集
 
 框架内置了对以下常见数据集格式的支持或提供了处理脚本：

{torch_rechub-0.1.0 → torch_rechub-0.3.0}/README_en.md

@@ -1,6 +1,8 @@
-# 🔥 Torch-RecHub - Lightweight, Efficient & Easy-to-use PyTorch Recommender Framework
+<div align="center">
 
-> 🚀 **30+ Mainstream Models** | 🎯 **Out-of-the-box** | 📦 **One-click ONNX Export** | 🤖 **Generative RecSys (HSTU/HLLM)**
+![Torch-RecHub Banner](docs/public/img/banner.png)
+
+# Torch-RecHub: A Lightweight, Efficient, and Easy-to-use PyTorch Recommender Framework
 
 [![License](https://img.shields.io/badge/license-MIT-blue?style=for-the-badge)](LICENSE)
 ![GitHub Repo stars](https://img.shields.io/github/stars/datawhalechina/torch-rechub?style=for-the-badge)
@@ -13,42 +15,33 @@
 [![numpy Version](https://img.shields.io/badge/numpy-1.19%2B-orange?style=for-the-badge)](https://numpy.org/)
 [![scikit-learn Version](https://img.shields.io/badge/scikit_learn-0.23%2B-orange?style=for-the-badge)](https://scikit-learn.org/)
 [![torch-rechub Version](https://img.shields.io/badge/torch_rechub-0.0.3%2B-orange?style=for-the-badge)](https://pypi.org/project/torch-rechub/)
+[![torchview](https://img.shields.io/badge/torchview-0.2%2B-green?style=for-the-badge)](https://github.com/mert-kurttutan/torchview)
 
 English | [简体中文](README.md)
 
-**Online Documentation:** https://datawhalechina.github.io/torch-rechub/ (English) | https://datawhalechina.github.io/torch-rechub/zh/ (简体中文)
+![Project Framework](docs/public/img/project_framework.png)
 
-**Torch-RecHub** — **Build production-grade recommender systems in 10 lines of code**. 30+ mainstream models out-of-the-box, one-click ONNX deployment, letting you focus on business instead of engineering.
+</div>
 
-![Torch-RecHub Banner](docs/public/img/banner.png)
+**Online Documentation:** https://datawhalechina.github.io/torch-rechub/
 
-## 🎯 Why Torch-RecHub?
-
-| Feature | Torch-RecHub | Other Frameworks |
-|---------|-------------|------------------|
-| Lines of Code | **10 lines** for train+eval+deploy | 100+ lines |
-| Model Coverage | **30+** mainstream models | Limited |
-| Generative RecSys | ✅ HSTU/HLLM (Meta 2024) | ❌ |
-| ONNX Export | ✅ Built-in support | Manual adaptation |
-| Learning Curve | Very Low | Steep |
+**Torch-RecHub** —— **Build production-grade recommender systems in 10 lines of code**. 30+ mainstream models out-of-the-box, one-click ONNX deployment, letting you focus on business instead of engineering.
 
 ## ✨ Features
 
 * **Modular Design:** Easy to add new models, datasets, and evaluation metrics.
-* **PyTorch-based:** Leverages PyTorch's dynamic graph and GPU acceleration capabilities.
-* **Rich Model Library:** Covers **30+** classic and cutting-edge recommendation algorithms (matching, ranking, multi-task, generative).
+* **Based on PyTorch:** Leverages PyTorch's dynamic graph and GPU acceleration capabilities.
+* **Rich Model Library:** Covers **30+** classic and cutting-edge recommendation algorithms (Matching, Ranking, Multi-task, Generative Recommendation, etc.).
 * **Standardized Pipeline:** Provides unified data loading, training, and evaluation workflows.
 * **Easy Configuration:** Adjust experiment settings via config files or command-line arguments.
 * **Reproducibility:** Designed to ensure reproducible experimental results.
-* **ONNX Export:** Export trained models to ONNX format for production deployment.
-* **Cross-engine data processing:** PySpark-based data processing and conversion supported for large-scale pipelines.
-* **Experiment visualization & tracking:** Unified integration of WandB, SwanLab, and TensorBoardX.
-* **Additional Features:** Negative sampling, multi-task learning, etc.
+* **ONNX Export:** Export trained models to ONNX format for seamless production deployment.
+* **Cross-engine Data Processing:** Support for PySpark-based data processing and transformation, facilitating deployment in big data pipelines.
+* **Experiment Visualization & Tracking:** Built-in unified integration for WandB, SwanLab, and TensorBoardX.
 
 ## 📖 Table of Contents
 
-- [🔥 Torch-RecHub - Lightweight, Efficient \& Easy-to-use PyTorch Recommender Framework](#-torch-rechub---lightweight-efficient--easy-to-use-pytorch-recommender-framework)
-  - [🎯 Why Torch-RecHub?](#-why-torch-rechub)
+- [🔥 Torch-RecHub - A Lightweight, Efficient, and Easy-to-use PyTorch Recommender Framework](#-torch-rechub---a-lightweight-efficient-and-easy-to-use-pytorch-recommender-framework)
   - [✨ Features](#-features)
   - [📖 Table of Contents](#-table-of-contents)
   - [🔧 Installation](#-installation)
@@ -61,7 +54,8 @@ English | [简体中文](README.md)
   - [🧪 Examples](#-examples)
     - [Ranking (CTR Prediction)](#ranking-ctr-prediction)
     - [Multi-Task Ranking](#multi-task-ranking)
-    - [Matching Model](#matching-model)
+    - [Matching Models](#matching-models)
+    - [Model Visualization](#model-visualization)
   - [👨‍💻‍ Contributors](#-contributors)
   - [🤝 Contributing](#-contributing)
   - [📜 License](#-license)
@@ -82,7 +76,7 @@ English | [简体中文](README.md)
 
 ### Installation Steps
 
-**Stable Version (Recommended for Users):**
+**Stable Version (Recommended):**
 ```bash
 pip install torch-rechub
 ```
@@ -147,7 +141,7 @@ torch-rechub/             # Root directory
 │   ├── matching/         # Matching task examples
 │   ├── ranking/          # Ranking task examples
 │   └── generative/       # Generative recommendation examples (HSTU, HLLM, etc.)
-├── docs/                 # Documentation (VitePress: multi-language, English & Chinese)
+├── docs/                 # Documentation (VitePress, multi-language)
 ├── tutorials/            # Jupyter tutorials
 ├── tests/                # Unit tests
 ├── config/               # Configuration files
@@ -158,54 +152,70 @@ torch-rechub/             # Root directory
 
 The framework currently supports **30+** mainstream recommendation models:
 
+<details>
+
 ### Ranking Models - 13
 
-| Model | Paper | Description |
-|-------|-------|-------------|
-| **DeepFM** | [IJCAI 2017](https://arxiv.org/abs/1703.04247) | FM + Deep joint training |
-| **Wide&Deep** | [DLRS 2016](https://arxiv.org/abs/1606.07792) | Memorization + Generalization |
-| **DCN** | [KDD 2017](https://arxiv.org/abs/1708.05123) | Explicit feature crossing |
-| **DCN-v2** | [WWW 2021](https://arxiv.org/abs/2008.13535) | Enhanced cross network |
-| **DIN** | [KDD 2018](https://arxiv.org/abs/1706.06978) | Attention for user interest |
-| **DIEN** | [AAAI 2019](https://arxiv.org/abs/1809.03672) | Interest evolution modeling |
-| **BST** | [DLP-KDD 2019](https://arxiv.org/abs/1905.06874) | Transformer for sequences |
-| **AFM** | [IJCAI 2017](https://arxiv.org/abs/1708.04617) | Attentional FM |
-| **AutoInt** | [CIKM 2019](https://arxiv.org/abs/1810.11921) | Auto feature interaction |
-| **FiBiNET** | [RecSys 2019](https://arxiv.org/abs/1905.09433) | Feature importance + Bilinear |
-| **DeepFFM** | [RecSys 2019](https://arxiv.org/abs/1611.00144) | Field-aware FM |
-| **EDCN** | [KDD 2021](https://arxiv.org/abs/2106.03032) | Enhanced DCN |
+| Model          | Paper                                             | Description                     |
+| -------------- | ------------------------------------------------- | ------------------------------- |
+| **DeepFM**    | [IJCAI 2017](https://arxiv.org/abs/1703.04247)   | FM + Deep joint training        |
+| **Wide&Deep** | [DLRS 2016](https://arxiv.org/abs/1606.07792)    | Memorization + Generalization   |
+| **DCN**       | [KDD 2017](https://arxiv.org/abs/1708.05123)     | Explicit feature crossing       |
+| **DCN-v2**    | [WWW 2021](https://arxiv.org/abs/2008.13535)     | Enhanced cross network          |
+| **DIN**       | [KDD 2018](https://arxiv.org/abs/1706.06978)     | Attention for user interest     |
+| **DIEN**      | [AAAI 2019](https://arxiv.org/abs/1809.03672)    | Interest evolution modeling     |
+| **BST**       | [DLP-KDD 2019](https://arxiv.org/abs/1905.06874) | Transformer for sequences       |
+| **AFM**       | [IJCAI 2017](https://arxiv.org/abs/1708.04617)   | Attentional FM                  |
+| **AutoInt**   | [CIKM 2019](https://arxiv.org/abs/1810.11921)    | Auto feature interaction learning |
+| **FiBiNET**   | [RecSys 2019](https://arxiv.org/abs/1905.09433)  | Feature importance + Bilinear   |
+| **DeepFFM**   | [RecSys 2019](https://arxiv.org/abs/1611.00144)  | Field-aware FM                  |
+| **EDCN**      | [KDD 2021](https://arxiv.org/abs/2106.03032)     | Enhanced DCN                    |
+
+</details>
+
+<details>
 
 ### Matching Models - 12
 
-| Model | Paper | Description |
-|-------|-------|-------------|
-| **DSSM** | [CIKM 2013](https://posenhuang.github.io/papers/cikm2013_DSSM_fullversion.pdf) | Classic two-tower model |
-| **YoutubeDNN** | [RecSys 2016](https://dl.acm.org/doi/10.1145/2959100.2959190) | YouTube deep retrieval |
-| **YoutubeSBC** | [RecSys 2019](https://dl.acm.org/doi/10.1145/3298689.3346997) | Sampling bias correction |
-| **MIND** | [CIKM 2019](https://arxiv.org/abs/1904.08030) | Multi-interest dynamic routing |
-| **SINE** | [WSDM 2021](https://arxiv.org/abs/2103.06920) | Sparse interest network |
-| **GRU4Rec** | [ICLR 2016](https://arxiv.org/abs/1511.06939) | GRU for sequences |
-| **SASRec** | [ICDM 2018](https://arxiv.org/abs/1808.09781) | Self-attentive sequential |
-| **NARM** | [CIKM 2017](https://arxiv.org/abs/1711.04725) | Neural attentive session |
-| **STAMP** | [KDD 2018](https://dl.acm.org/doi/10.1145/3219819.3219895) | Short-term attention memory |
-| **ComiRec** | [KDD 2020](https://arxiv.org/abs/2005.09347) | Controllable multi-interest |
+| Model           | Paper                                                                           | Description               |
+| --------------- | ------------------------------------------------------------------------------- | ------------------------- |
+| **DSSM**       | [CIKM 2013](https://posenhuang.github.io/papers/cikm2013_DSSM_fullversion.pdf) | Classic two-tower model   |
+| **YoutubeDNN** | [RecSys 2016](https://dl.acm.org/doi/10.1145/2959100.2959190)                  | YouTube deep retrieval    |
+| **YoutubeSBC** | [RecSys 2019](https://dl.acm.org/doi/10.1145/3298689.3346997)                  | Sampling bias correction  |
+| **MIND**       | [CIKM 2019](https://arxiv.org/abs/1904.08030)                                  | Multi-interest dynamic routing |
+| **SINE**       | [WSDM 2021](https://arxiv.org/abs/2103.06920)                                  | Sparse interest network   |
+| **GRU4Rec**    | [ICLR 2016](https://arxiv.org/abs/1511.06939)                                  | GRU for sequences         |
+| **SASRec**     | [ICDM 2018](https://arxiv.org/abs/1808.09781)                                  | Self-attentive sequential |
+| **NARM**       | [CIKM 2017](https://arxiv.org/abs/1711.04725)                                  | Neural attentive session  |
+| **STAMP**      | [KDD 2018](https://dl.acm.org/doi/10.1145/3219819.3219895)                     | Short-term attention memory priority |
+| **ComiRec**    | [KDD 2020](https://arxiv.org/abs/2005.09347)                                   | Controllable multi-interest |
+
+</details>
+
+<details>
 
 ### Multi-Task Models - 5
 
-| Model | Paper | Description |
-|-------|-------|-------------|
-| **ESMM** | [SIGIR 2018](https://arxiv.org/abs/1804.07931) | Entire space multi-task |
-| **MMoE** | [KDD 2018](https://dl.acm.org/doi/10.1145/3219819.3220007) | Multi-gate mixture-of-experts |
-| **PLE** | [RecSys 2020](https://dl.acm.org/doi/10.1145/3383313.3412236) | Progressive layered extraction |
-| **AITM** | [KDD 2021](https://arxiv.org/abs/2105.08489) | Adaptive information transfer |
-| **SharedBottom** | - | Classic shared bottom |
+| Model             | Paper                                                          | Description               |
+| ----------------- | -------------------------------------------------------------- | ------------------------- |
+| **ESMM**         | [SIGIR 2018](https://arxiv.org/abs/1804.07931)                | Entire space multi-task   |
+| **MMoE**         | [KDD 2018](https://dl.acm.org/doi/10.1145/3219819.3220007)    | Multi-gate Mixture-of-Experts |
+| **PLE**          | [RecSys 2020](https://dl.acm.org/doi/10.1145/3383313.3412236) | Progressive Layered Extraction |
+| **AITM**         | [KDD 2021](https://arxiv.org/abs/2105.08489)                  | Adaptive Information Transfer |
+| **SharedBottom** | -                                                             | Classic shared bottom     |
+
+</details>
+
+<details>
 
 ### Generative Recommendation - 2
 
-| Model | Paper | Description |
-|-------|-------|-------------|
+| Model     | Paper                                          | Description                                              |
+| --------- | ---------------------------------------------- | -------------------------------------------------------- |
 | **HSTU** | [Meta 2024](https://arxiv.org/abs/2402.17152) | Hierarchical Sequential Transduction Units, powering Meta's trillion-parameter RecSys |
-| **HLLM** | [2024](https://arxiv.org/abs/2409.12740) | Hierarchical LLM for recommendation, combining LLM semantic understanding |
+| **HLLM** | [2024](https://arxiv.org/abs/2409.12740)      | Hierarchical LLM for recommendation, combining LLM semantic understanding |
+
+</details>
 
 ## 📊 Supported Datasets
 
@@ -219,7 +229,7 @@ The framework provides built-in support or preprocessing scripts for the followi
 * **BookCrossing**
 * **Ali-ccp**
 * **Yidian**
-* ...
+* ... 
 
 The expected data format is typically an interaction file containing:
 - User ID
@@ -231,7 +241,6 @@ For specific format requirements, please refer to the example code in the `tutor
 
 You can easily integrate your own datasets by ensuring they conform to the framework's data format requirements or by writing custom data loaders.
 
-
 ## 🧪 Examples
 
 All model usage examples can be found in `/examples`
@@ -260,7 +269,7 @@ ctr_trainer.export_onnx("deepfm.onnx")
 from torch_rechub.models.multi_task import SharedBottom, ESMM, MMOE, PLE, AITM
 from torch_rechub.trainers import MTLTrainer
 
-task_types = ["classification", "classification"]
+task_types = ["classification", "classification"] 
 model = MMOE(features, task_types, 8, expert_params={"dims": [32,16]}, tower_params_list=[{"dims": [32, 16]}, {"dims": [32, 16]}])
 
 mtl_trainer = MTLTrainer(model)
@@ -269,7 +278,7 @@ auc = ctr_trainer.evaluate(ctr_trainer.model, test_dataloader)
 mtl_trainer.export_onnx("mmoe.onnx")
 ```
 
-### Matching Model
+### Matching Models
 
 ```python
 from torch_rechub.models.matching import DSSM
@@ -282,19 +291,27 @@ train_dl, test_dl, item_dl = dg.generate_dataloader(test_user, all_item, batch_s
 model = DSSM(user_features, item_features, temperature=0.02,
              user_params={
                  "dims": [256, 128, 64],
-                 "activation": 'prelu',
+                 "activation": 'prelu',  
              },
              item_params={
                  "dims": [256, 128, 64],
-                 "activation": 'prelu',
+                 "activation": 'prelu', 
              })
 
 match_trainer = MatchTrainer(model)
 match_trainer.fit(train_dl)
 match_trainer.export_onnx("dssm.onnx")
-# For two-tower models, you can also export user and item towers separately:
-# match_trainer.export_onnx("dssm_user.onnx", tower="user")
-# match_trainer.export_onnx("dssm_item.onnx", tower="item")
+# For two-tower models, you can export user and item towers separately:
+# match_trainer.export_onnx("user_tower.onnx", mode="user")
+# match_trainer.export_onnx("item_tower.onnx", mode="item")
+```
+
+### Model Visualization
+
+```python
+# Visualize model architecture (Requires: pip install torch-rechub[visualization])
+graph = ctr_trainer.visualization(depth=4)  # Generate computation graph
+ctr_trainer.visualization(save_path="model.pdf", dpi=300)  # Save as high-resolution PDF
 ```
 
 ## 👨‍💻‍ Contributors
@@ -333,11 +350,9 @@ If you use this framework in your research or work, please consider citing:
 
 ## 📫 Contact
 
-* **Project Lead:** [1985312383](https://github.com/1985312383)
+* **Project Lead:** [1985312383](https://github.com/1985312383) 
 * [**GitHub Discussions**](https://github.com/datawhalechina/torch-rechub/discussions)
 
-
-
 ## ⭐️ Star History
 
 [![Star History Chart](https://api.star-history.com/svg?repos=datawhalechina/torch-rechub&type=Date)](https://www.star-history.com/#datawhalechina/torch-rechub&Date)

torch_rechub-0.3.0/docs/en/community/changelog.md

@@ -0,0 +1,9 @@
+---
+title: Changelog
+description: Torch-RecHub version update history
+---
+
+# Changelog
+
+This page is under construction.
+