search

uta-smile / DeepAttnMISL

Contains codes of DeepAttnMISL in MEDIA'20
55 stars 19 forks source link

Keras Implementation of the Code #1

Open narminGhaffari opened 3 years ago

narminGhaffari commented 3 years ago

Thank you for sharing this code. I am trying to implement this code. However, I have the main question regarding:

lbl_pred = model(graph, masked_cls)

In this line of code, the graph contains the images and maskled_cls is the cluster labels, right? the output of this model, is binary values of 0 and 1, indicating the event value or it is the survival time?

eballot21 commented 3 years ago

Dear Narmin Ghaffari,

I am trying to implement the code too in my slides but I got an error : The object named 'graph' is a torch.cuda.DoubleTensor and the object masked_cls is a torch.cuda.FloatTensor and should be in the same format.

Yao Jiawen advised me to make it as the float32 tensor.

Could you please confirm that the expected format for the "graph" object is a Float Tensor?

Thank you in advance Elise