Thanks for the great work and code! I can’t match the paper’s metrics on my setup. Could you share the exact training command/config you used (optimizer + LR/schedule, epochs, batch size, loss weights, augmentations/padding, seed, eval script/settings)?
Thanks for the great work and code! I can’t match the paper’s metrics on my setup. Could you share the exact training command/config you used (optimizer + LR/schedule, epochs, batch size, loss weights, augmentations/padding, seed, eval script/settings)?