Open Jaclye opened 5 months ago
Yes, the training dataset have a shape of [N_records, 3, duration * sampling_frequency], the 3
means E, N, Z
components. You should keep such order for your training dataset. The model automatically learns such order when you use the trained model for generating waveforms
Hi author, does filtering or baseline correction come first in data processing?
Yes, preprocessing methods are needed. A high quality dataset is crucial for training the model. Please refer to closed question #4 and question #5 for more information about preprocessing pipeline
Hello, author. I see that the order in which the images are generated is E N Z. I should enter them in that order when making the numpy array right?