Commit 48d93b2

committed

multihead_attention: optimize shapes

1 parent 741f200 commit 48d93b2Copy full SHA for 48d93b2

2 files changed

+173

-163

lines changed

src/nf
- nf_multihead_attention.f90
test
- test_multihead_attention_layer.f90

2 files changed

+173

-163

lines changed

Comments

(0)