NCF

PyTorch Implementation of "Neural Collaborative Filtering" at WWW'17

I used the ml-100k dataset.

1. General Framework

Concatenate user and item embeddings, then pass through Neural CF Layers to predict score

2. Fusion of GMF and MLP (NeuMF)

GMF that applies a linear kernel to model the latent feature interactions, and MLP that uses a non-linear kernel to learn the interaction function from data

Fuse GMF and MLP under the NCF framework

3. Requirements

numpy==1.24.1
pandas==1.5.2
scikit_learn==1.2.1
scikit_surprise==1.1.3
surprise==0.1
torch==1.13.1

4. Example run

set config.json

{
  "seed": 417,
  "batch_size": 4096,
  "learning_rate": 0.001,
  "weight_decay": 0.01,
  "test_size": 0.2,
  "epochs": 20,
  "save_dir": "./save_models",
  "arch": {
    "embedding_size": 128,
    "mlp_layer_dims": [64, 32, 16],
    "dropout_rate": 0.2,
    "use_gmf": true
  }
}

run python code

python train.py --config [config.json file path]

ryubright/NCF

NCF

1. General Framework

2. Fusion of GMF and MLP (NeuMF)

3. Requirements

4. Example run