ML project: implementation of research paper
Budget: ₹600 – ₹1,500 INR
Implement paper: Xu, N., Zhang, H., Liu, A. A., Nie, W., Su, Y., Nie, J., & Zhang, Y. (2019). Multi-level policy and reward-based deep reinforcement learning
framework for image captioning. IEEE Transactions on Multimedia, 22(5), 1372-1383.
Do it in google colab.
Dataset to be used MSCOCO (and flickr optional).
Focus should be on reaching the accuracy as close to given if get higher accuracy then huge plus but not required.
There are few approach given by authors. cnn-rnn approach of the paper is must and one of ours-stacked method(2 given)
ours-stacked method have bleu1 of 80 in implemented should reach atleast 76 or near.
cnn-rnn method have bleu1 of 75 should reach atleast 71 or near.
evaluation metric bleu1.
Can use code from other paper implementation (but need to mention all the places from where the code has been used) .
framework for image captioning. IEEE Transactions on Multimedia, 22(5), 1372-1383.
Do it in google colab.
Dataset to be used MSCOCO (and flickr optional).
Focus should be on reaching the accuracy as close to given if get higher accuracy then huge plus but not required.
There are few approach given by authors. cnn-rnn approach of the paper is must and one of ours-stacked method(2 given)
ours-stacked method have bleu1 of 80 in implemented should reach atleast 76 or near.
cnn-rnn method have bleu1 of 75 should reach atleast 71 or near.
evaluation metric bleu1.
Can use code from other paper implementation (but need to mention all the places from where the code has been used) .