blacksnail789521

Same question here

@hrzn Or maybe you can show me that how I should implement my custom `dataset` to achieve `stride` and `group`. It seems that we need to modify the class as...

> **Describe potential alternatives** > Right now I'd have to implement my own `TrainingDataset` and `InferenceDataset` to achieve this custom sampling functionality. @maxkazmsft Can you share your code of `TrainingDataset`...

``` from causalnex.structure.notears import from_pandas from causalnex.structure.pytorch.notears import from_pandas as from_pandas_pytorch ```

Sorry, but I'm currently working on my papers for my Ph.D. program. I'll try to work on this repo after finishing the papers :)

It's the timestamp information. https://github.com/zhouhaoyi/Informer2020/blob/0ac81e04d4095ecb97a3a78c7b49c936d8aa9933/data/data_loader.py#L176-L177 Also, it seems that a lot of people confuse "mark" with "mask". (https://github.com/zhouhaoyi/Informer2020/issues/158#issuecomment-848691957)

Yes, it used the [`time_features`](https://github.com/zhouhaoyi/Informer2020/blob/0ac81e04d4095ecb97a3a78c7b49c936d8aa9933/utils/timefeatures.py#L114) to create the timestamp infos.

My solution is to store one trial into one folder. So, if I had 16 trials, there would be 16 folders. ``` hp_path_for_this_trial = os.path.join(hp_path, trial_id) model.fit( x_train, y_train, epochs=1,...