make TransformerLayer accept a bshd or sbhd tensor format
#2003
| Job | Run time |
|---|---|
| 0s | |
| 0s | |
| 0s | |
| 0s |
bshd or sbhd tensor format
#2003
| Job | Run time |
|---|---|
| 0s | |
| 0s | |
| 0s | |
| 0s |