Questions of reproducing BTCV

[mt-cly]

Thanks for your open-source work.
I try to reproduce the BTCV performance based on your code, python train.py -net sam -mod sam_adpt -exp_name msa-3d-sam-btcv -sam_ckpt ./checkpoint/sam/sam_vit_b_01ec64.pth -image_size 192 -b 8 -dataset decathlon -thd True -chunk 96 -evl_chunk 96 -data_path ./data/ -num_sample 1 but the DICE results are unstable and much worse than paper report (~0.4 v.s., 0.883). Have you any suggestions or can you share the config or command reproducing the performance in BTCV. Thanks.

[JaireYu]

Hi @mt-cly, please refer to the pull request #135, it fixes few bugs and may help.

[l412198735]

Hi @mt-cly , have you resolved the issue?