I am working with bigger dataset (coco)
but the problem is that after training with the same model as in the video lectures, after some epochs my error starts increasing at it never comes down in the later epochs.
I cant find out the problem here(maybe with the learning rate)
Training of caption bot model
Hey,
That’s totally experimental conclusions which architecture we should choose.
Since you data is larger than this one. I would suggest you to make your model more complex, by adding more layers. Maybe this can help to improve the accuracy
I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.
On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.
can you please solve this error, i checked my code and i do not find any error but it keeps on showing this error while training the model
In the data_generator function on line no. 10 remove ‘.jpg’. So line 10 will look like this
photo = encoding_train[key]
Hii thanks for the help i fixed it. But i need to ask one thing that my model is not that much accurate so what should i do to increase its accuracy
In the current model, only lstm and cnn’s are used. There are some other things also like Bi-LSTMs. which can help in reducing the loss, you can further stack them to build a more accurate model.
Also there are some more things like transformers and attention model which can help to improve the model further
1 Like