ML4ITS/mtad-gat-pytorch
PyTorch implementation of MTAD-GAT (Multivariate Time-Series Anomaly Detection via Graph Attention Networks) by Zhao et. al (2020, https://arxiv.org/abs/2009.02040).
PythonMIT
Pinned issues
Issues
- 4
tqdm is not imported in prediction.py + request
#7 opened by TKrom - 0
Running it on custome dataset with 20 features
#44 opened by Avv22 - 2
about adjust_predicts() ,please!!!
#40 opened by fffii - 0
Versions of Ptyhon and Pytorch
#43 opened by zyz760543032 - 0
Evaluation Code
#42 opened by LameloBally - 1
Computational resource
#32 opened by khadijakhaldi - 0
- 1
Embedding vector dimension issue in the paper
#39 opened by 0nyoun9 - 1
about the example output,please!!!
#38 opened by goodman233 - 1
Multiple inconsistent training results
#37 opened by ZhangYN1226 - 1
The issue with the dataset
#33 opened by Dudududu15 - 2
- 1
- 0
Pot results on the SMD dataset
#35 opened by wangxuekui123 - 2
- 2
My understanding of mtad _ gat.py does not reflect the ' Multivariate Time-series ' in the title of the paper.
#31 opened by yuanyuanyuantang - 2
some question about model and the result
#10 opened by 2snoopy88 - 1
- 1
The reason why use shuffle in time-series data
#26 opened by cloudhs7 - 0
about gat_layer
#30 opened by adverbial03 - 1
ValueError: time data '<built-in function id>' does not match format '%d%m%Y_%H%M%S'
#17 opened by KANGWOOLEE1234 - 1
Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same
#23 opened by xiaokai01 - 1
FileNotFoundError: [Errno 2] No such file or directory: 'datasets/ServerMachineDataset/processed\\machine-1-1_train.pkl'
#22 opened by xiaokai01 - 1
About data cleaning
#20 opened by hariiiseldon - 1
Some question about the param target_dims
#11 opened by ShaoSK - 2
FC layer out_dim not matching RECOn layer in_dim
#18 opened by krenusz - 1
Running repo on custom data
#15 opened by hiddensquid1409 - 1
Reason for using `find_epsilon` in feature-level
#14 opened by cloudhs7 - 2
When computing 'e' in the FeatureAttentionLayer, the output of MTAD_GAT 'predictions', 'recons' are all nan, and training is not possible due to the presence of nan.
#13 opened by ylic204 - 7
losses are always nan
#12 opened by m-ali-awan - 3
A Question about the implementation.
#2 opened by Kevin-XiongC - 1
attention layer
#5 opened by wjj5881005 - 2
out_dim or n_features
#3 opened by ZhixuanLiu - 1
The purpose of `adjust_anomaly_scores`
#4 opened by severous - 2
The parameter of `adjust_predicts()`
#6 opened by vvvu