List your company
ResearchFormerly WizWand

GPU Memory Usage on Hecktor 2022 Medical Image Segmentation

562Peak GPU Memory Usage (Training)Slim UNETR
-230.45,118.310,46715,815.7Sep 26, 2025
Updated 4mo ago

Evaluation results

MethodLinks
Slim UNETRType=CNN-Transformer, Batch size=22025.09562124
VeloxSegType=CNN-Transformer, Batch size=22025.091,006178
U-KANType=CNN-KAN, Batch size=22025.092,360640
SegFormer-3DType=CNN-Transformer, Batch size=22025.092,702436
H-DenseFormerType=CNN-Transformer, Batch size=22025.093,7781,474
UNETR++Type=CNN-Transformer, Batch size=22025.093,9061,088
VNetType=CNN, Batch size=22025.094,4602,072
UNETRType=CNN-Transformer, Batch size=22025.094,6261,914
NestedFormerType=CNN-Transformer, Batch size=22025.095,1821,966
U-RWKVType=CNN-RWKV, Batch size=22025.095,3881,044
UNetType=CNN, Batch size=22025.095,9422,698
HCMA-UNetType=CNN-Mamba, Batch size=22025.099,0981,892
A2FSegType=CNN, Batch size=22025.099,1022,352
SuperLightNetType=CNN-Transformer, Batch size=22025.099,1922,730
Swin UNETRType=CNN-Transformer, Batch size=22025.0913,2583,704
VSmTransType=CNN-Transformer, Batch size=22025.0913,3182,532
MedNeXtType=CNN, Batch size=22025.0920,3724,376