MultiHeadAttention._compute_attention_mask()
always returns a bool …
#35
actions.yml
on: push
Check the code format
2m 32s
Matrix: Run tests