Ambiguity in evaluation results on MNIST digit classification

davidtellez / contrastive-predictive-coding

Keras implementation of Representation Learning with Contrastive Predictive Coding

525 stars 120 forks source link

Ambiguity in evaluation results on MNIST digit classification #14

Open bhavinjawade opened 3 years ago

bhavinjawade commented 3 years ago

Hi David,

Its a great repo. I have doubt regarding the final 90% accuracy that you received on the MNSIT dataset in 10 epochs. Isn't that necessarily because of the moderately deep MLP that you used. Typically an network that deep easily gives a 90% accuracy on MNIST in few epochs. I am not sure myself, so thought to ask.

Thanks and regards, Bhavin Jawade