nextrec 0.4.3

A comprehensive recommendation library with match, ranking, and multi-task learning models

English | 中文文档

A Unified, Efficient, and Scalable Recommendation System Framework

Table of Contents

• Introduction
• Installation
• Architecture
• 5-Minute Quick Start
• CLI Usage
• Platform Compatibility
• Supported Models
• Contributing

Introduction

NextRec is a modern recommendation system framework built on PyTorch, providing researchers and engineering teams with a fast modeling, training, and evaluation experience. The framework adopts a modular design with rich built-in model implementations, data processing tools, and engineering-ready training components, covering various recommendation scenarios. NextRec provides easy-to-use interfaces, command-line tools, and tutorials, enabling recommendation algorithm learners to quickly understand model architectures and train and infer models at the fastest speed.

Why NextRec

• Unified feature engineering & data pipeline: NextRec provides Dense/Sparse/Sequence feature definitions, persistent DataProcessor, and batch-optimized RecDataLoader, matching the model training and inference process based on offline parquet/csv features in industrial big-data Spark/Hive scenarios.
• Multi-scenario recommendation capabilities: Covers ranking (CTR/CVR), retrieval, multi-task learning and other recommendation/marketing models, with a continuously expanding model zoo.
• Developer-friendly experience: Supports stream preprocessing/distributed training/inference for various data formats (csv/parquet/pathlike), GPU acceleration and visual metric monitoring, facilitating experiments for business algorithm engineers and recommendation algorithm learners.
• Flexible command-line tool: Through configuring training and inference config files, start training and inference processes with one command nextrec --mode=train --train_config=train_config.yaml, facilitating rapid experiment iteration and agile deployment.
• Efficient training & evaluation: NextRec's standardized training engine comes with various optimizers, learning rate schedulers, early stopping, model checkpoints, and detailed log management built-in, ready to use out of the box.

Architecture

NextRec adopts a modular and low-coupling engineering design, enabling full-pipeline reusability and scalability across data processing → model construction → training & evaluation → inference & deployment. Its core components include: a Feature-Spec-driven Embedding architecture, the BaseModel abstraction, a set of independent reusable Layers, a unified DataLoader for both training and inference, and a ready-to-use Model Zoo.

The project borrows ideas from excellent open-source rec libraries. Early layers referenced torch-rechub but have been replaced with in-house implementations. torch-rechub remains mature in architecture and models; the author contributed a bit there—feel free to check it out.

---

Installation

You can quickly install the latest NextRec via pip install nextrec; Python 3.10+ is required.

Tutorials

We provide multiple examples in the tutorials/ directory, covering ranking, retrieval, multi-task, and data processing scenarios:

• movielen_ranking_deepfm.py — DeepFM model training example on MovieLens 100k dataset
• example_ranking_din.py — DIN deep interest network training example on e-commerce dataset
• example_multitask.py — ESMM multi-task learning training example on e-commerce dataset
• movielen_match_dssm.py — DSSM retrieval model example trained on MovieLens 100k dataset
• run_all_ranking_models.py — Quickly verify the availability of all ranking models
• run_all_multitask_models.py — Quickly verify the availability of all multi-task models
• run_all_match_models.py — Quickly verify the availability of all retrieval models

If you want to learn more details about the NextRec framework, we also provide Jupyter notebooks to help you understand:

• How to get started with the NextRec framework
• How to use the data processor for data preprocessing

5-Minute Quick Start

We provide a detailed quick start guide and paired datasets to help you become familiar with different features of the NextRec framework. We provide a test dataset from an e-commerce scenario in the datasets/ path, with data examples as follows:

user_id	item_id	dense_0	dense_1	dense_2	dense_3	dense_4	dense_5	dense_6	dense_7	sparse_0	sparse_1	sparse_2	sparse_3	sparse_4	sparse_5	sparse_6	sparse_7	sparse_8	sparse_9	sequence_0	sequence_1	label
1	7817	0.14704075	0.31020382	0.77780896	0.944897	0.62315375	0.57124174	0.77009535	0.3211029	315	260	379	146	168	161	138	88	5	312	[170,175,97,338,105,353,272,546,175,545,463,128,0,0,0]	[368,414,820,405,548,63,327,0,0,0,0,0,0,0,0]	0
1	3579	0.77811223	0.80359334	0.5185201	0.91091245	0.043562356	0.82142705	0.8803686	0.33748195	149	229	442	6	167	252	25	402	7	168	[179,48,61,551,284,165,344,151,0,0,0,0,0,0,0]	[814,0,0,0,0,0,0,0,0,0,0,0,0,0,0]	1

Next, we'll use a short example to show you how to train a DIN model using NextRec. DIN (Deep Interest Network) is from Alibaba's 2018 KDD Best Paper, used for CTR prediction scenarios. You can also directly execute python tutorials/example_ranking_din.py to run the training and inference code.

After starting training, you can view detailed training logs in the nextrec_logs/din_tutorial path.

import pandas as pd

from nextrec.models.ranking.din import DIN
from nextrec.basic.features import DenseFeature, SparseFeature, SequenceFeature

df = pd.read_csv('dataset/ranking_task.csv')

for col in df.columns and 'sequence' in col: # csv loads lists as text; convert them back to objects
    df[col] = df[col].apply(lambda x: eval(x) if isinstance(x, str) else x)

# Define feature columns
dense_features = [DenseFeature(name=f'dense_{i}', input_dim=1) for i in range(8)]

sparse_features = [SparseFeature(name='user_id', embedding_name='user_emb', vocab_size=int(df['user_id'].max() + 1), embedding_dim=32), SparseFeature(name='item_id', embedding_name='item_emb', vocab_size=int(df['item_id'].max() + 1), embedding_dim=32),]

sparse_features.extend([SparseFeature(name=f'sparse_{i}', embedding_name=f'sparse_{i}_emb', vocab_size=int(df[f'sparse_{i}'].max() + 1), embedding_dim=32) for i in range(10)])

sequence_features = [
    SequenceFeature(name='sequence_0', vocab_size=int(df['sequence_0'].apply(lambda x: max(x)).max() + 1), embedding_dim=32, padding_idx=0, embedding_name='item_emb'),
    SequenceFeature(name='sequence_1', vocab_size=int(df['sequence_1'].apply(lambda x: max(x)).max() + 1), embedding_dim=16, padding_idx=0, embedding_name='sparse_0_emb'),]

mlp_params = {
    "dims": [256, 128, 64],
    "activation": "relu",
    "dropout": 0.3,
}

model = DIN(
    dense_features=dense_features,
    sparse_features=sparse_features,
    sequence_features=sequence_features,
    mlp_params=mlp_params,
    attention_hidden_units=[80, 40],
    attention_activation='sigmoid',
    attention_use_softmax=True,
    target=['label'],                                     # target variable
    device='mps',                                         
    embedding_l1_reg=1e-6,
    embedding_l2_reg=1e-5,
    dense_l1_reg=1e-5,
    dense_l2_reg=1e-4,
    session_id="din_tutorial",                            # experiment id for logs
)

# Compile model with optimizer and loss
model.compile(
            optimizer = "adam",
            optimizer_params = {"lr": 1e-3, "weight_decay": 1e-5},
            loss = "focal",
            loss_params={"gamma": 2.0, "alpha": 0.25},
        )

model.fit(
    train_data=df,
    metrics=['auc', 'gauc', 'logloss'],  # metrics to track
    epochs=3,
    batch_size=512,
    shuffle=True,
    user_id_column='user_id'             # used for GAUC
)

# Evaluate after training
metrics = model.evaluate(
    df,
    metrics=['auc', 'gauc', 'logloss'],
    batch_size=512,
    user_id_column='user_id'
)

CLI Usage

NextRec provides a powerful command-line interface for model training and prediction using YAML configuration files. For detailed CLI documentation, see:

• NextRec CLI User Guide - Complete guide for using the CLI

# Train a model
nextrec --mode=train --train_config=path/to/train_config.yaml

# Run prediction
nextrec --mode=predict --predict_config=path/to/predict_config.yaml

As of version 0.4.3, NextRec CLI supports single-machine training; distributed training features are currently under development.

Platform Compatibility

The current version is 0.4.3. All models and test code have been validated on the following platforms. If you encounter compatibility issues, please report them in the issue tracker with your system version:

Platform	Configuration
MacOS latest	MacBook Pro M4 Pro 24GB RAM
Ubuntu latest	AutoDL 4070D Dual GPU
CentOS 7	Intel Xeon 5138Y 96 cores 377GB RAM

---

Supported Models

Ranking Models

Model	Paper	Year	Status
FM	Factorization Machines	ICDM 2010	Supported
AFM	Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks	IJCAI 2017	Supported
DeepFM	DeepFM: A Factorization-Machine based Neural Network for CTR Prediction	IJCAI 2017	Supported
Wide&Deep	Wide & Deep Learning for Recommender Systems	DLRS 2016	Supported
xDeepFM	xDeepFM: Combining Explicit and Implicit Feature Interactions	KDD 2018	Supported
FiBiNET	FiBiNET: Combining Feature Importance and Bilinear Feature Interaction for CTR Prediction	RecSys 2019	Supported
PNN	Product-based Neural Networks for User Response Prediction	ICDM 2016	Supported
AutoInt	AutoInt: Automatic Feature Interaction Learning	CIKM 2019	Supported
DCN	Deep & Cross Network for Ad Click Predictions	ADKDD 2017	Supported
DCN v2	DCN V2: Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systems	KDD 2021	In Progress
DIN	Deep Interest Network for CTR Prediction	KDD 2018	Supported
DIEN	Deep Interest Evolution Network	AAAI 2019	Supported
MaskNet	MaskNet: Feature-wise Gating Blocks for High-dimensional Sparse Recommendation Data	2020	Supported

Retrieval Models

Model	Paper	Year	Status
DSSM	Learning Deep Structured Semantic Models	CIKM 2013	Supported
DSSM v2	DSSM with pairwise BPR-style optimization	-	Supported
YouTube DNN	Deep Neural Networks for YouTube Recommendations	RecSys 2016	Supported
MIND	Multi-Interest Network with Dynamic Routing	CIKM 2019	Supported
SDM	Sequential Deep Matching Model	-	Supported

Multi-task Models

Model	Paper	Year	Status
MMOE	Modeling Task Relationships in Multi-task Learning	KDD 2018	Supported
PLE	Progressive Layered Extraction	RecSys 2020	Supported
ESMM	Entire Space Multi-task Model	SIGIR 2018	Supported
ShareBottom	Multitask Learning	-	Supported
POSO	POSO: Personalized Cold-start Modules for Large-scale Recommender Systems	2021	Supported

Generative Models

Model	Paper	Year	Status
TIGER	Recommender Systems with Generative Retrieval	NeurIPS 2023	In Progress
HSTU	Hierarchical Sequential Transduction Units	-	Supported

---

Contributing

We welcome contributions of any form!

How to Contribute

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add AmazingFeature')
4. Push your branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Before submitting a PR, please run python test/run_tests.py and python scripts/format_code.py to ensure all tests pass and code style is unified.

Code Style

• Follow PEP8
• Provide unit tests for new functionality
• Update documentation accordingly

Reporting Issues

When submitting issues on GitHub, please include:

• Description of the problem
• Reproduction steps
• Expected behavior
• Actual behavior
• Environment info (Python version, PyTorch version, etc.)

---

License

This project is licensed under the Apache 2.0 License.

---

Contact

• GitHub Issues: Submit an issue
• Email: [e-mail]

---

Acknowledgements

NextRec is inspired by the following great open-source projects:

• torch-rechub — Flexible, easy-to-extend recommendation framework
• FuxiCTR — Configurable, tunable, and reproducible CTR library
• RecBole — Unified, comprehensive, and efficient recommendation library

Special thanks to all open-source contributors!

---

Back to Top