GPU
Harzva opened this issue · comments
I found that this task does not use multiple gpu training ,why? thank you very much.
Using our provided hyper-parameters, the whole model can be trained on a Tesla V100 GPU with 32GB memory
Official implementation of the TMM paper "Relation-aware Compositional Zero-shot Learning for Attribute-Object Pair Recognition".
Harzva opened this issue · comments
I found that this task does not use multiple gpu training ,why? thank you very much.
Using our provided hyper-parameters, the whole model can be trained on a Tesla V100 GPU with 32GB memory