This is the pytorch implement of Matching Networks for One Shot Learning

Train

If you want to train the model,simply run the code

python mainOmniglot.py 

You can set fce = True if you want use Full Context Embeddings.You can use the follow set in mainOmniglot.py for a 5-way one shot learning.

classes_per_set = 5
samples_per_class = 1

Result

After about 30 epoches,you can achieve about 97% accuracy on train set and 96% on val and test set.

Acknowledgement

This work refer to gitabcworld's implement,detail athttps://github.com/AntreasAntoniou/MatchingNetworks.