Increased inference throughput to reflect lower inference sampling rate
Switching log normal parameters for evaluation to better reflect GWTC-3
I have not generated a test dataset using the new SNR threshold, so the trained model will need to be evaluated on that. But have otherwise confirmed this reproduces SOA results and is ready for production.
I have not generated a test dataset using the new SNR threshold, so the trained model will need to be evaluated on that. But have otherwise confirmed this reproduces SOA results and is ready for production.