popcornell/keras-triplet-center-loss

Simple Keras implementation of Triplet-Center Loss on the MNIST dataset

Hi there,

I'm a Post-Doc at CMU LTI, working on Audio Processing and Machine Learning.

You can find a link to my CV here.

---

keras-triplet-center-loss's Issues

The distance between the sample and its nearest negative center is not computed at all.

As described in the title, it seems that np.inf is assigned to the intra-class distances in triplet_center_loss in triplet.py, which is more like some kind of placeholder.

triplet_center_loss:TypeError

when i want to run the triplet_center_loss, the error is show that ,how can i fix it, thanks.

"TypeError: Input 'y' of 'Equal' Op has type float32 that does not match type uint8 of argument 'x'"

TypeError: unhashable type: 'numpy.ndarray'

Hi!!
when generate_triplet function is called I am getting an error. Please help me to resolve it
this is the line for this error
for data_class in sorted(set(data_xy[1])):

reshape error

x_train, x_test, y_train, y_test = train_test_split(images, labels, test_size=0.33, shuffle= True)

model_input = Input(shape=(224,224, 3))
softmax, pre_logits = base_network()
x_train_flat = x_train.reshape(-1, 163968)

x_train_flat = x_train.reshape(-1, 163968)
ValueError: cannot reshape array of size 1304626176 into shape (163968)

what is the exact reason for this. please help me