Closed suhangpro closed 8 years ago
Either numerical instability in the softmax layer because np.linalg.norm(caffe2tf_1d_blob('prob') - tf_activations['prob']
is a bit larger than the distance of other layers. And/or from the randomness in the dropout layers
fixed in 97a9e1ff663ac7238f986d8b8247f4ab81f83432
Any idea why this happens? Awesome work by the way!