This is a code for URP research project I have been doing for 6 months (Jan - July 2017). The project title was "Developing new mechanisms to Neural Turing Machines (NTMs) for one-shot learning".
Python 100.00%
meta-learning-via-feature-label-memory-network's People
1.What '5th' mean? Does it mean the image from some class was seen for the fifth time?
2.Would we clear the memory after one episode and use a new memory in the new episode?
3.What we do during test time? Would we update the memory just as in the training time?
Thank you very much!