Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
[Model][Hetero] GCMC using new hetero APIs (#860)
* init * rm data * README * fix rating values that are not decimal * rm stale codes * small fix * upd * rewrite decoder * fix many * many fix; performance matched * upd; handle sparse input * upd * address comments * more docstring; download data automatically * shared param mode
- Loading branch information
1 parent
e3921d5
commit 9737b27
Showing
6 changed files
with
1,135 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1 @@ | ||
log |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,51 @@ | ||
# Graph Convolutional Matrix Completion | ||
|
||
Paper link: [https://arxiv.org/abs/1706.02263](https://arxiv.org/abs/1706.02263) | ||
Author's code: [https://github.com/riannevdberg/gc-mc](https://github.com/riannevdberg/gc-mc) | ||
|
||
The implementation does not handle side-channel features and mini-epoching and thus achieves | ||
slightly worse performance when using node features. | ||
|
||
Credit: Jiani Zhang ([@jennyzhang0215](https://github.com/jennyzhang0215)) | ||
|
||
## Dependencies | ||
* MXNet 1.5.0+ | ||
* pandas | ||
* gluonnlp | ||
|
||
## Data | ||
|
||
Supported datasets: ml-100k, ml-1m, ml-10m | ||
|
||
## How to run | ||
|
||
ml-100k, no feature | ||
```bash | ||
DGLBACKEND=mxnet python train.py --data_name=ml-100k --use_one_hot_fea --gcn_agg_accum=stack | ||
``` | ||
Results: RMSE=0.9077 (0.910 reported) | ||
Speed: 0.0246s/epoch (vanilla implementation: 0.1008s/epoch) | ||
|
||
ml-100k, with feature | ||
```bash | ||
DGLBACKEND=mxnet python train.py --data_name=ml-100k --gcn_agg_accum=stack | ||
``` | ||
Results: RMSE=0.9495 (0.905 reported) | ||
|
||
ml-1m, no feature | ||
```bash | ||
DGLBACKEND=mxnet python train.py --data_name=ml-1m --gcn_agg_accum=sum --use_one_hot_fea | ||
``` | ||
Results: RMSE=0.8377 (0.832 reported) | ||
Speed: 0.0695s/epoch (vanilla implementation: 1.538s/epoch) | ||
|
||
ml-10m, no feature | ||
```bash | ||
DGLBACKEND=mxnet python train.py --data_name=ml-10m --gcn_agg_accum=stack --gcn_dropout=0.3 \ | ||
--train_lr=0.001 --train_min_lr=0.0001 --train_max_iter=15000 \ | ||
--use_one_hot_fea --gen_r_num_basis_func=4 | ||
``` | ||
Results: RMSE=0.7875 (0.777 reported) | ||
Speed: 0.6480s/epoch (vanilla implementation: OOM) | ||
|
||
Testbed: EC2 p3.2xlarge |
Oops, something went wrong.