K-Means is EM
Gibbs Sampling is EM
OMG, VAE and GAN are also EM
OMG, the gradient is exploding!
Chill, SGD will save us all.
That moment when you realize cv2 and tf.image give you totally different results 🙄
An AlexNet finetune report in Tensorflow
Yes, I still play with ancient NNs
OMG, how could implementation of Padding and Lrn be different between Tensorflow and Caffe
Yeah, positive results in the end
But, but it’s not really solved.