WEN Hao

I think I know the reason now. The [CPSC2020 dataset](https://github.com/DeepPSP/torch_ecg/blob/master/benchmarks/train_hybrid_cpsc2020/dataset.py) uses sliced recordings since the original recordings are fairly long. So, you should call the [persistence](https://github.com/DeepPSP/torch_ecg/blob/821dfd8064be8e03aae46ef872a341013d4867c8/benchmarks/train_hybrid_cpsc2020/dataset.py#L414-L436) method first, which takes...

I added automatic downloading of these models, which you can find in https://opensz.oss-cn-beijing.aliyuncs.com/ICBEB2020/file/CPSC2019-opensource.zip. However, these models were trained with a very older version of Keras. One might have trouble loading...

The reason for this warning seems to be that you are not using the latest version, since the code ```python ampl = torch.as_tensor(ampl, dtype=sig.dtype, device=sig.device).unsqueeze(-1) ``` where this warning is...

Was the error raised when using `torch_ecg.databases.CPSC2021` or `torch_ecg.databases.datasets.CPSC2021Dataset`? Currently when passing empty `db_dir`, a warning instead of an error should be raised, for example ```python from torch_ecg.databases import CPSC2021...

I might have found the problem. One has to set `db_dir` in the config class instance for `torch_ecg.databases.datasets.CPSC2021Dataset`, which is `None` by default.

`ED` is for [`easydict.EasyDict`](https://pypi.org/project/easydict/) which was previously used as a configuration class, and has already been replaced with `torch_ecg.cfg.CFG`. I plan to replace `torch_ecg.cfg.CFG` with `dataclass` because it still has...

Peaks are almost the "argmax" of corresponding waves, perhaps you can do a peak detection (sometimes valley detection) within the ranges of the waves after wave delineation (segmentation). Or you...

> 您好，我是否可以单独使用您的preprocessors来给我自己的数据集去噪、切分心拍？ `preprocessors` can be used for denoising. As for slicing the heartbeats, you can use traditional methods like [`wfdb.processing.XQRS`](https://wfdb.readthedocs.io/en/latest/processing.html?highlight=xqrs#wfdb.processing.XQRS) or some DL model, for example, [models from CPSC2019](https://github.com/DeepPSP/torch_ecg/tree/master/benchmarks/train_multi_cpsc2019).

OK, I will check this soon.

I tested it with the recording `HR06004` from [sample-data/cinc2021](https://github.com/DeepPSP/torch_ecg/tree/master/sample-data/cinc2021), and it worked normally:  Could you please provide me your `raw_sig`?