Amharic Language Image Captions Generation Using Hybridized Attention-Based Deep Neural Networks

Figure 5

Performance of the model on the BNATURE dataset. (a) Loss on train and test. (b) Train and test accuracy.

(a)

(b)