LogADEmpirical icon indicating copy to clipboard operation
LogADEmpirical copied to clipboard

CNN and NeuralLog cannot be trained properly

Open Timkeeper2018 opened this issue 9 months ago • 0 comments

When I use cnn.yaml and neurallog.yaml for training, the following error occurs.

04/24/2025 10:24:44 - INFO - CNN - Train data statistics: {'min': 2, 'max': 278, 'mean': np.float64(19.401958056550622)}
Sliding window with size 10: 100%|██████████| 51756/51756 [00:03<00:00, 15243.86it/s]
04/24/2025 10:24:48 - INFO - CNN - Number of sequences: 542693
04/24/2025 10:24:48 - INFO - CNN - Number of normal sequence: 536102
04/24/2025 10:24:48 - INFO - CNN - Number of abnormal sequence: 6591
04/24/2025 10:24:48 - WARNING - CNN - Number of unique abnormal events: 0
04/24/2025 10:24:48 - INFO - CNN - Number of abnormal sessions: 1481/51756
Traceback (most recent call last):
  File "D:\LogADEmpirical\main_run.py", line 396, in <module>
    run(args)
  File "D:\LogADEmpirical\main_run.py", line 375, in run
    train_and_eval(args,
  File "D:\LogADEmpirical\main_run.py", line 201, in train_and_eval
    logger.info(f"Train dataset: {len(sequentials)}")
                                  ^^^^^^^^^^^^^^^^
TypeError: object of type 'NoneType' has no len()

How can I solve this problem? Thanks

Timkeeper2018 avatar Apr 24 '25 02:04 Timkeeper2018