Afsanay

Why are you not using positional encoding in the encoder of the network??

Can you please provide us with the idea on how were you able to generate real-time captions on test images?