Closed 110rangeeeee closed 3 months ago
Hi, @110rangeeeee
Due to poor hardware, I only used part CODE 15% in this implementation. In this repository, I only use those signals with no arrhythmia and crop all signals to 7 seconds. You can access the dataset via the OneDrive link on the main page of this repository. The shape for npz files is (N, 12, 4096). For more details about CODE 15%, you shall check info via this link.
Hello, we can no longer use OneDrive here. Could you please share the link using a different method?
Hello, we can no longer use OneDrive here. Could you please share the link using a different method?
Oh, I don't have another Cloud Drive with enough space for the dataset. You can still get those data via zenodo, though they use HDF5 format. Please check details about the dataset via zenodo website.
You can load those hdf5 file with following code:
import h5py
r = h5py.File(f_name.hdf5, 'r')
idx, ecg = np.array(r['exam_id']),np.array(r['tracings'])
Hello, could you please share some information about the data read from target_signals_dir? Specifically regarding files ending in .npz. Thank you.