Pretrained weights on Stanford datasets
Closed this issue · 2 comments
SIYUAN9446 commented
Thanks for sharing the code of such interesting work.
Could you please also provide the pre-trained weights on Stanford 2D FSE and Stanford 3D FSE datasets?
Thanks in advance!
z-fabian commented
Hi @SIYUAN9446,
Thanks for reaching out. We uploaded the Stanford checkpoints and you can find them here.
Please keep in mind that there is no train/test split defined on these datasets, so we trained 3 models from scratch on different random train-val splits and averaged the results in the paper. The uploaded checkpoints correspond to setting train_val_seed
to 0
and train_val_split
to 0.8
.
SIYUAN9446 commented
Hi @z-fabian,
Thanks a lot for your reply.