2-dor

Been training some batch models in the latest version & couldn't help but notice the recent feature to create a nam model after each epoch eats up a bit of...

[BUG] Pre-training validation throws error in v0.10.0 for 3 reamps (batch training of 111 reamps)

2

Hi Steve, I've been doing batch training for a larger pack of 111 reamps. No issues (that I can see) in the signal chain or with the reamped files but...

[FEATURE] LR_DECAY adjustment from 0.007 to 0.0045 seems to help model convergence

8

Hey Steve, Minor thing but it seems that using the LR_DECAY value of 0.0045 in the trainer benefits model convergence over the 0.007 value in the trainer by default. I...

[BUG] Final prediction ESR doesn't match what's computed during the trainer's validation step

1

Hi Steve, I've been re-training some models with trainer version v0.10.0 and encountered something I know happened a while back. In the "checkpoints" folder, my lowest ESR seems to have...