Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
yxy235 fedaa36da5 | 6 months ago | |
---|---|---|
.. | ||
README.md | 1 year ago | |
builder.py | 1 year ago | |
data_utils.py | 6 months ago | |
evaluation.py | 1 year ago | |
layers.py | 1 year ago | |
model.py | 1 year ago | |
model_sparse.py | 1 year ago | |
process_movielens1m.py | 9 months ago | |
process_nowplaying_rs.py | 1 year ago | |
sampler.py | 1 year ago |
python process_movielens1m.py ./ml-1m ./data_processed
.ml-1m
with the directory you put the .dat
files, and replace data_processed
withpython process_nowplaying_rs.py ./nowplaying_rs_dataset ./data_processed
This model returns items that are K nearest neighbors of the latest item the user has
interacted. The distance between two items are measured by Euclidean distance of
item embeddings, which are learned as outputs of PinSAGE.
python model.py data_processed --num-epochs 300 --num-workers 2 --device cuda:0 --hidden-dims 64
The implementation here also assigns a learnable vector to each item. If your hidden
state size is so large that the learnable vectors cannot fit into GPU, use this script
for sparse embedding update (written with torch.optim.SparseAdam
) instead:
python model_sparse.py data_processed --num-epochs 300 --num-workers 2 --device cuda:0 --hidden-dims 1024
Note that since the embedding update is done on CPU, it will be significantly slower than doing
everything on GPU.
The HITS@10 is 0.01241, compared to 0.01220 with SLIM with the same dimensionality.\
The implementation here is different from what being described in the paper:
No Description
Python C++ Jupyter Notebook Cuda Text other
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》