Official Implementation of FRec

SIGIR'24 Paper: Modeling User Fatigue for Sequential Recommendation

This project is based on Microsoft Recommender and was originally designed for TensorFlow 2.1.
However, to run the model on modern hardware (e.g., RTX 30xx GPUs), we recommend upgrading to TensorFlow 2.4.1 due to compatibility with CUDA 11.0 and cuDNN 8.x.

---

🔧 Environment Setup (Python 3.8 + TensorFlow 2.4.1)

✅ 1. Install Python 3.8.10

Download from: https://www.python.org/downloads/release/python-3810/

✅ 2. Create a virtual environment

& "C:\Users\<your_username>\AppData\Local\Programs\Python\Python38\python.exe" -m venv frec-tf24-env
.frec-tf24-env\Scripts\activate

✅ 3. Install dependencies

Create a file requirements_tf24.txt with:

numpy
pandas
scikit-learn
PyYAML
tqdm
absl-py~=0.10
setproctitle==1.1.10
tensorflow==2.4.1
protobuf==3.19.6
retrying==1.3.3

Then run:

pip install -r requirements_tf24.txt

⚠️ Make sure you have installed CUDA 11.0 and cuDNN 8.x.

---

🧪 Running the Model

python run.py --model model --name trial

Other baselines:

python run.py --model clsr --name trial
python run.py --model slirec --name trial

---

🛠 Common Fixes

If you see ModuleNotFoundError: No module named 'sli_rec', add this at the top of run.py:

import sys, os
sys.path.append(os.path.dirname(os.path.abspath(__file__)))

To run on CPU only:

import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"

---

🗑 Deactivate and Remove Virtual Environment

To deactivate:

deactivate

To delete the virtual environment (PowerShell):

Remove-Item -Recurse -Force .\frec-tf24-env

Or manually delete the folder frec-tf24-env from File Explorer.

---

📁 Data

We provide a processed Taobao dataset. To extract:

unzip data.zip

Data format follows standard sequential input from the Microsoft Recommender.

---

📌 Important Hyperparameters

• num_cross_layers: number of cross layers
• recent_k: truncated sequence length
• num_interests: number of latent interests
• k_size: convolutional kernel size
• alpha: contrastive learning weight